Search

Stakeholders -- from model developers to policymakers -- seek to minimize the dual-use risks of large language models (LLMs). An open challenge to this goal is whether technical safeguards can impede the misuse of LLMs, even when models are customizable via fine-tuning or when model weights are fully open. In response, several recent studies have proposed methods to produce durable LLM safeguards for open-weight LLMs that can withstand adversarial modifications of the model's weights via fine-tuning. This holds the promise of raising adversaries' costs even under strong threat models where adversaries can directly fine-tune model weights. However, in this paper, we urge for more careful characterization of the limits of these approaches. Through several case studies, we demonstrate that even evaluating these defenses is exceedingly difficult and can easily mislead audiences into thinking that safeguards are more durable than they really are. We draw lessons from the evaluation pitfalls that we identify and suggest future research carefully cabin claims to more constrained, well-defined, and rigorously examined threat models, which can provide more useful and candid assessments to stakeholders.

Artificial intelligence (AI) model creators commonly attach restrictive terms of use to both their models and their outputs. These terms typically prohibit activities ranging from creating competing AI models to spreading disinformation. Often taken at face value, these terms are positioned by companies as key enforceable tools for preventing misuse, particularly in policy dialogs. But are these terms truly meaningful? There are myriad examples where these broad terms are regularly and repeatedly violated. Yet except for some account suspensions on platforms, no model creator has actually tried to enforce these terms with monetary penalties or injunctive relief. This is likely for good reason: we think that the legal enforceability of these licenses is questionable. This Article systematically assesses of the enforceability of AI model terms of use and offers three contributions. First, we pinpoint a key problem: the artifacts that they protect, namely model weights and model outputs, are largely not copyrightable, making it unclear whether there is even anything to be licensed. Second, we examine the problems this creates for other enforcement. Recent doctrinal trends in copyright preemption may further undermine state-law claims, while other legal frameworks like the DMCA and CFAA offer limited recourse. Anti-competitive provisions likely fare even worse than responsible use provisions. Third, we provide recommendations to policymakers. There are compelling reasons for many provisions to be unenforceable: they chill good faith research, constrain competition, and create quasi-copyright ownership where none should exist. There are, of course, downsides: model creators have fewer tools to prevent harmful misuse. But we think the better approach is for statutory provisions, not private fiat, to distinguish between good and bad uses of AI, restricting the latter., Comment: Forthcoming Indiana Law Journal

Large language models are finetuned to refuse questions about hazardous knowledge, but these protections can often be bypassed. Unlearning methods aim at completely removing hazardous capabilities from models and make them inaccessible to adversaries. This work challenges the fundamental differences between unlearning and traditional safety post-training from an adversarial perspective. We demonstrate that existing jailbreak methods, previously reported as ineffective against unlearning, can be successful when applied carefully. Furthermore, we develop a variety of adaptive methods that recover most supposedly unlearned capabilities. For instance, we show that finetuning on 10 unrelated examples or removing specific directions in the activation space can recover most hazardous capabilities for models edited with RMU, a state-of-the-art unlearning method. Our findings challenge the robustness of current unlearning approaches and question their advantages over safety training., Comment: Spotlight paper at Neurips 2024 SoLaR workshop

Language models (LMs) derive their capabilities from extensive training on diverse data, including potentially copyrighted material. These models can memorize and generate content similar to their training data, posing potential concerns. Therefore, model creators are motivated to develop mitigation methods that prevent generating protected content. We term this procedure as copyright takedowns for LMs, noting the conceptual similarity to (but legal distinction from) the DMCA takedown This paper introduces the first evaluation of the feasibility and side effects of copyright takedowns for LMs. We propose CoTaEval, an evaluation framework to assess the effectiveness of copyright takedown methods, the impact on the model's ability to retain uncopyrightable factual knowledge from the training data whose recitation is embargoed, and how well the model maintains its general utility and efficiency. We examine several strategies, including adding system prompts, decoding-time filtering interventions, and unlearning approaches. Our findings indicate that no tested method excels across all metrics, showing significant room for research in this unique problem setting and indicating potential unresolved challenges for live policy proposals., Comment: 31 pages, 9 figures, 14 tables

Foundation model development attracts a rapidly expanding body of contributors, scientists, and applications. To help shape responsible development practices, we introduce the Foundation Model Development Cheatsheet: a growing collection of 250+ tools and resources spanning text, vision, and speech modalities. We draw on a large body of prior work to survey resources (e.g. software, documentation, frameworks, guides, and practical tools) that support informed data selection, processing, and understanding, precise and limitation-aware artifact documentation, efficient model training, advance awareness of the environmental impact from training, careful model evaluation of capabilities, risks, and claims, as well as responsible model release, licensing and deployment practices. We hope this curated collection of resources helps guide more responsible development. The process of curating this list, enabled us to review the AI development ecosystem, revealing what tools are critically missing, misused, or over-used in existing practices. We find that (i) tools for data sourcing, model evaluation, and monitoring are critically under-serving ethical and real-world needs, (ii) evaluations for model safety, capabilities, and environmental impact all lack reproducibility and transparency, (iii) text and particularly English-centric analyses continue to dominate over multilingual and multi-modal analyses, and (iv) evaluation of systems, rather than just models, is needed so that capabilities and impact are assessed in context.

Evaluating aligned large language models' (LLMs) ability to recognize and reject unsafe user requests is crucial for safe, policy-compliant deployments. Existing evaluation efforts, however, face three limitations that we address with SORRY-Bench, our proposed benchmark. First, existing methods often use coarse-grained taxonomies of unsafe topics, and are over-representing some fine-grained topics. For example, among the ten existing datasets that we evaluated, tests for refusals of self-harm instructions are over 3x less represented than tests for fraudulent activities. SORRY-Bench improves on this by using a fine-grained taxonomy of 45 potentially unsafe topics, and 450 class-balanced unsafe instructions, compiled through human-in-the-loop methods. Second, linguistic characteristics and formatting of prompts are often overlooked, like different languages, dialects, and more -- which are only implicitly considered in many evaluations. We supplement SORRY-Bench with 20 diverse linguistic augmentations to systematically examine these effects. Third, existing evaluations rely on large LLMs (e.g., GPT-4) for evaluation, which can be computationally expensive. We investigate design choices for creating a fast, accurate automated safety evaluator. By collecting 7K+ human annotations and conducting a meta-evaluation of diverse LLM-as-a-judge designs, we show that fine-tuned 7B LLMs can achieve accuracy comparable to GPT-4 scale LLMs, with lower computational cost. Putting these together, we evaluate over 40 proprietary and open-source LLMs on SORRY-Bench, analyzing their distinctive refusal behaviors. We hope our effort provides a building block for systematic evaluations of LLMs' safety refusal capabilities, in a balanced, granular, and efficient manner.

Recent studies show that image and video generation models can be prompted to reproduce copyrighted content from their training data, raising serious legal concerns around copyright infringement. Copyrighted characters, in particular, pose a difficult challenge for image generation services, with at least one lawsuit already awarding damages based on the generation of these characters. Yet, little research has empirically examined this issue. We conduct a systematic evaluation to fill this gap. First, we build CopyCat, an evaluation suite consisting of diverse copyrighted characters and a novel evaluation pipeline. Our evaluation considers both the detection of similarity to copyrighted characters and generated image's consistency with user input. Our evaluation systematically shows that both image and video generation models can still generate characters even if characters' names are not explicitly mentioned in the prompt, sometimes with only two generic keywords (e.g., prompting with "videogame, plumber" consistently generates Nintendo's Mario character). We then introduce techniques to semi-automatically identify such keywords or descriptions that trigger character generation. Using our evaluation suite, we study runtime mitigation strategies, including both existing methods and new strategies we propose. Our findings reveal that commonly employed strategies, such as prompt rewriting in the DALL-E system, are not sufficient as standalone guardrails. These strategies must be coupled with other approaches, like negative prompting, to effectively reduce the unintended generation of copyrighted characters. Our work provides empirical grounding to the discussion of copyright mitigation strategies and offers actionable insights for model deployers actively implementing them.

The safety alignment of current Large Language Models (LLMs) is vulnerable. Relatively simple attacks, or even benign fine-tuning, can jailbreak aligned models. We argue that many of these vulnerabilities are related to a shared underlying issue: safety alignment can take shortcuts, wherein the alignment adapts a model's generative distribution primarily over only its very first few output tokens. We refer to this issue as shallow safety alignment. In this paper, we present case studies to explain why shallow safety alignment can exist and provide evidence that current aligned LLMs are subject to this issue. We also show how these findings help explain multiple recently discovered vulnerabilities in LLMs, including the susceptibility to adversarial suffix attacks, prefilling attacks, decoding parameter attacks, and fine-tuning attacks. Importantly, we discuss how this consolidated notion of shallow safety alignment sheds light on promising research directions for mitigating these vulnerabilities. For instance, we show that deepening the safety alignment beyond just the first few tokens can often meaningfully improve robustness against some common exploits. Finally, we design a regularized finetuning objective that makes the safety alignment more persistent against fine-tuning attacks by constraining updates on initial tokens. Overall, we advocate that future safety alignment should be made more than just a few tokens deep.

In this study, we investigate the vulnerability of image watermarks to diffusion-model-based image editing, a challenge exacerbated by the computational cost of accessing gradient information and the closed-source nature of many diffusion models. To address this issue, we introduce JIGMARK. This first-of-its-kind watermarking technique enhances robustness through contrastive learning with pairs of images, processed and unprocessed by diffusion models, without needing a direct backpropagation of the diffusion process. Our evaluation reveals that JIGMARK significantly surpasses existing watermarking solutions in resilience to diffusion-model edits, demonstrating a True Positive Rate more than triple that of leading baselines at a 1% False Positive Rate while preserving image quality. At the same time, it consistently improves the robustness against other conventional perturbations (like JPEG, blurring, etc.) and malicious watermark attacks over the state-of-the-art, often by a large margin. Furthermore, we propose the Human Aligned Variation (HAV) score, a new metric that surpasses traditional similarity measures in quantifying the number of image derivatives from image editing.

The exposure of security vulnerabilities in safety-aligned language models, e.g., susceptibility to adversarial attacks, has shed light on the intricate interplay between AI safety and AI security. Although the two disciplines now come together under the overarching goal of AI risk management, they have historically evolved separately, giving rise to differing perspectives. Therefore, in this paper, we advocate that stakeholders in AI risk management should be aware of the nuances, synergies, and interplay between safety and security, and unambiguously take into account the perspectives of both disciplines in order to devise mostly effective and holistic risk mitigation approaches. Unfortunately, this vision is often obfuscated, as the definitions of the basic concepts of "safety" and "security" themselves are often inconsistent and lack consensus across communities. With AI risk management being increasingly cross-disciplinary, this issue is particularly salient. In light of this conceptual challenge, we introduce a unified reference framework to clarify the differences and interplay between AI safety and AI security, aiming to facilitate a shared understanding and effective collaboration across communities.

Instruction tuning is an important step in making language models useful for direct user interaction. However, many legal tasks remain out of reach for most open LLMs and there do not yet exist any large scale instruction datasets for the domain. This critically limits research in this application area. In this work, we curate LawInstruct, a large legal instruction dataset, covering 17 jurisdictions, 24 languages and a total of 12M examples. We present evidence that domain-specific pretraining and instruction tuning improve performance on LegalBench, including improving Flan-T5 XL by 8 points or 16\% over the baseline. However, the effect does not generalize across all tasks, training regimes, model sizes, and other factors. LawInstruct is a resource for accelerating the development of models with stronger information processing and decision making capabilities in the legal domain.

Current Large Language Models (LLMs), even those tuned for safety and alignment, are susceptible to jailbreaking. Some have found that just further fine-tuning an aligned model with benign data (i.e., data without harmful content) surprisingly leads to substantial degradation in safety. We delve into the data-centric aspects of why benign fine-tuning inadvertently contributes to jailbreaking. First, we represent fine-tuning data through two lenses: representation and gradient spaces. Additionally, we propose a bi-directional anchoring method that, during the selection process, prioritizes data points that are close to harmful examples and far from benign ones. Our approach effectively identifies subsets of benign data that are more likely to degrade the model's safety after fine-tuning. Training on just 100 of these seemingly benign datapoints surprisingly leads to the fine-tuned model affirmatively responding to >70% of tested harmful requests, compared to <20% after fine-tuning on randomly selected data. We also observe that the selected data frequently appear as lists, bullet points, or math questions, indicating a systematic pattern in fine-tuning data that contributes to jailbreaking.

Independent evaluation and red teaming are critical for identifying the risks posed by generative AI systems. However, the terms of service and enforcement strategies used by prominent AI companies to deter model misuse have disincentives on good faith safety evaluations. This causes some researchers to fear that conducting such research or releasing their findings will result in account suspensions or legal reprisal. Although some companies offer researcher access programs, they are an inadequate substitute for independent research access, as they have limited community representation, receive inadequate funding, and lack independence from corporate incentives. We propose that major AI developers commit to providing a legal and technical safe harbor, indemnifying public interest safety research and protecting it from the threat of account suspensions or legal reprisal. These proposals emerged from our collective experience conducting safety, privacy, and trustworthiness research on generative AI systems, where norms and incentives could be better aligned with public interests, without exacerbating model misuse. We believe these commitments are a necessary step towards more inclusive and unimpeded community efforts to tackle the risks of generative AI.

Foundation models are powerful technologies: how they are released publicly directly shapes their societal impact. In this position paper, we focus on open foundation models, defined here as those with broadly available model weights (e.g. Llama 2, Stable Diffusion XL). We identify five distinctive properties (e.g. greater customizability, poor monitoring) of open foundation models that lead to both their benefits and risks. Open foundation models present significant benefits, with some caveats, that span innovation, competition, the distribution of decision-making power, and transparency. To understand their risks of misuse, we design a risk assessment framework for analyzing their marginal risk. Across several misuse vectors (e.g. cyberattacks, bioweapons), we find that current research is insufficient to effectively characterize the marginal risk of open foundation models relative to pre-existing technologies. The framework helps explain why the marginal risk is low in some cases, clarifies disagreements about misuse risks by revealing that past work has focused on different subsets of the framework with different assumptions, and articulates a way forward for more constructive debate. Overall, our work helps support a more grounded assessment of the societal impact of open foundation models by outlining what research is needed to empirically validate their theoretical benefits and risks.

Large language models (LLMs) show inherent brittleness in their safety mechanisms, as evidenced by their susceptibility to jailbreaking and even non-malicious fine-tuning. This study explores this brittleness of safety alignment by leveraging pruning and low-rank modifications. We develop methods to identify critical regions that are vital for safety guardrails, and that are disentangled from utility-relevant regions at both the neuron and rank levels. Surprisingly, the isolated regions we find are sparse, comprising about $3\%$ at the parameter level and $2.5\%$ at the rank level. Removing these regions compromises safety without significantly impacting utility, corroborating the inherent brittleness of the model's safety mechanisms. Moreover, we show that LLMs remain vulnerable to low-cost fine-tuning attacks even when modifications to the safety-critical regions are restricted. These findings underscore the urgent need for more robust safety strategies in LLMs., Comment: 22 pages, 9 figures. Project page is available at https://boyiwei.com/alignment-attribution/

Is AI set to redefine the legal profession? We argue that this claim is not supported by the current evidence. We dive into AI's increasingly prevalent roles in three types of legal tasks: information processing; tasks involving creativity, reasoning, or judgment; and predictions about the future. We find that the ease of evaluating legal applications varies greatly across legal tasks, based on the ease of identifying correct answers and the observability of information relevant to the task at hand. Tasks that would lead to the most significant changes to the legal profession are also the ones most prone to overoptimism about AI capabilities, as they are harder to evaluate. We make recommendations for better evaluation and deployment of AI in legal contexts.

Optimizing large language models (LLMs) for downstream use cases often involves the customization of pre-trained LLMs through further fine-tuning. Meta's open release of Llama models and OpenAI's APIs for fine-tuning GPT-3.5 Turbo on custom datasets also encourage this practice. But, what are the safety costs associated with such custom fine-tuning? We note that while existing safety alignment infrastructures can restrict harmful behaviors of LLMs at inference time, they do not cover safety risks when fine-tuning privileges are extended to end-users. Our red teaming studies find that the safety alignment of LLMs can be compromised by fine-tuning with only a few adversarially designed training examples. For instance, we jailbreak GPT-3.5 Turbo's safety guardrails by fine-tuning it on only 10 such examples at a cost of less than $0.20 via OpenAI's APIs, making the model responsive to nearly any harmful instructions. Disconcertingly, our research also reveals that, even without malicious intent, simply fine-tuning with benign and commonly used datasets can also inadvertently degrade the safety alignment of LLMs, though to a lesser extent. These findings suggest that fine-tuning aligned LLMs introduces new safety risks that current safety infrastructures fall short of addressing -- even if a model's initial safety alignment is impeccable, it is not necessarily to be maintained after custom fine-tuning. We outline and critically analyze potential mitigations and advocate for further research efforts toward reinforcing safety protocols for the custom fine-tuning of aligned LLMs.

This guidance report focuses on improving the quality of teaching of mathematics to pupils in Key Stages 2 and 3. The decision to focus on these Key Stages was made after an initial consultation period with teachers, academics, and other stakeholders. Schools are seeking advice on adjusting to a new curriculum, and there is also concern about pupils making a transition between the stages. This report is not intended to provide a comprehensive guide to mathematics teaching, but rather is an accessible overview of existing research with clear, actionable guidance. Recommendations based on research findings that schools can use to make a significant difference to pupils' learning, and those based on the questions that appear to be most salient to practitioners, are provided. Those recommendations include: (1) Use assessment to build upon pupils' existing knowledge and understanding; (2) Use manipulatives and representations; (3) Teach strategies for solving problems; (4) Enable pupils to develop a rich network of mathematical knowledge; (5) Develop pupils' independence and motivation; (6) Use tasks and resources to challenge and support pupils' mathematics; (7) Use structured interventions to provide additional support; and (8) Support pupils to make a successful transition between primary and secondary school.

The advent of large language models (LLMs) and their adoption by the legal community has given rise to the question: what types of legal reasoning can LLMs perform? To enable greater study of this question, we present LegalBench: a collaboratively constructed legal reasoning benchmark consisting of 162 tasks covering six different types of legal reasoning. LegalBench was built through an interdisciplinary process, in which we collected tasks designed and hand-crafted by legal professionals. Because these subject matter experts took a leading role in construction, tasks either measure legal reasoning capabilities that are practically useful, or measure reasoning skills that lawyers find interesting. To enable cross-disciplinary conversations about LLMs in the law, we additionally show how popular legal frameworks for describing legal reasoning -- which distinguish between its many forms -- correspond to LegalBench tasks, thus giving lawyers and LLM developers a common vocabulary. This paper describes LegalBench, presents an empirical evaluation of 20 open-source and commercial LLMs, and illustrates the types of research explorations LegalBench enables., Comment: 143 pages, 79 tables, 4 figures

Is the output of generative AI entitled to First Amendment protection? We're inclined to say yes. Even though current AI programs are of course not people and do not themselves have constitutional rights, their speech may potentially be protected because of the rights of the programs' creators. But beyond that, and likely more significantly, AI programs' speech should be protected because of the rights of their users-both the users' rights to listen and their rights to speak. In this short Article, we sketch the outlines of this analysis., Comment: Published in the Journal of Free Speech Law (2023)

Generative AI, in particular text-based "foundation models" (large models trained on a huge variety of information including the internet), can generate speech that could be problematic under a wide range of liability regimes. Machine learning practitioners regularly "red team" models to identify and mitigate such problematic speech: from "hallucinations" falsely accusing people of serious misconduct to recipes for constructing an atomic bomb. A key question is whether these red-teamed behaviors actually present any liability risk for model creators and deployers under U.S. law, incentivizing investments in safety mechanisms. We examine three liability regimes, tying them to common examples of red-teamed model behaviors: defamation, speech integral to criminal conduct, and wrongful death. We find that any Section 230 immunity analysis or downstream liability analysis is intimately wrapped up in the technical details of algorithm design. And there are many roadblocks to truly finding models (and their associated parties) liable for generated speech. We argue that AI should not be categorically immune from liability in these scenarios and that as courts grapple with the already fine-grained complexities of platform algorithms, the technical details of generative AI loom above with thornier questions. Courts and policymakers should think carefully about what technical design incentives they create as they evaluate these issues., Comment: Published in the Journal of Free Speech Law (2023)

Recently, there has been a surge of interest in integrating vision into Large Language Models (LLMs), exemplified by Visual Language Models (VLMs) such as Flamingo and GPT-4. This paper sheds light on the security and safety implications of this trend. First, we underscore that the continuous and high-dimensional nature of the visual input makes it a weak link against adversarial attacks, representing an expanded attack surface of vision-integrated LLMs. Second, we highlight that the versatility of LLMs also presents visual attackers with a wider array of achievable adversarial objectives, extending the implications of security failures beyond mere misclassification. As an illustration, we present a case study in which we exploit visual adversarial examples to circumvent the safety guardrail of aligned LLMs with integrated vision. Intriguingly, we discover that a single visual adversarial example can universally jailbreak an aligned LLM, compelling it to heed a wide range of harmful instructions that it otherwise would not) and generate harmful content that transcends the narrow scope of a `few-shot' derogatory corpus initially employed to optimize the adversarial example. Our study underscores the escalating adversarial risks associated with the pursuit of multimodality. Our findings also connect the long-studied adversarial vulnerabilities of neural networks to the nascent field of AI alignment. The presented attack suggests a fundamental adversarial challenge for AI alignment, especially in light of the emerging trend toward multimodality in frontier foundation models.

Large language models (LLMs) power many state-of-the-art systems in natural language processing. However, these models are extremely computationally expensive, even at inference time, raising the natural question: when is the extra cost of deploying a larger model worth the anticipated boost in capabilities? Better understanding this tradeoff fundamentally could benefit from an inference efficiency metric that is both (i) easily comparable across models from different providers, and (ii) representative of the true cost of running queries in an isolated performance environment. Unfortunately, access to LLMs today is largely restricted to black-box text generation APIs and raw runtimes measured through this interface do not satisfy these desiderata: model providers can apply various software and hardware optimizations orthogonal to the model, and models served on shared infrastructure are susceptible to performance contention. To circumvent these problems, we propose a new metric for comparing inference efficiency across models. This metric puts models on equal footing as though they were served (i) on uniform hardware and software, and (ii) without performance contention. We call this metric the \emph{idealized runtime}, and we propose a methodology to efficiently estimate this metric for autoregressive Transformer models. We also propose cost-aware variants that incorporate the number of accelerators needed to serve the model. Using these metrics, we compare ten state-of-the-art LLMs to provide the first analysis of inference efficiency-capability tradeoffs; we make several observations from this analysis, including the fact that the superior inference runtime performance of certain APIs is often a byproduct of optimizations within the API rather than the underlying model. Our methodology also facilitates the efficient comparison of different software and hardware stacks.

Existing foundation models are trained on copyrighted material. Deploying these models can pose both legal and ethical risks when data creators fail to receive appropriate attribution or compensation. In the United States and several other countries, copyrighted content may be used to build foundation models without incurring liability due to the fair use doctrine. However, there is a caveat: If the model produces output that is similar to copyrighted data, particularly in scenarios that affect the market of that data, fair use may no longer apply to the output of the model. In this work, we emphasize that fair use is not guaranteed, and additional work may be necessary to keep model development and deployment squarely in the realm of fair use. First, we survey the potential risks of developing and deploying foundation models based on copyrighted content. We review relevant U.S. case law, drawing parallels to existing and potential applications for generating text, source code, and visual art. Experiments confirm that popular foundation models can generate content considerably similar to copyrighted material. Second, we discuss technical mitigations that can help foundation models stay in line with fair use. We argue that more research is needed to align mitigation strategies with the current state of the law. Lastly, we suggest that the law and technical mitigations should co-evolve. For example, coupled with other policy mechanisms, the law could more explicitly consider safe harbors when strong technical tools are used to mitigate infringement harms. This co-evolution may help strike a balance between intellectual property and innovation, which speaks to the original goal of fair use. But we emphasize that the strategies we describe here are not a panacea and more work is needed to develop policies that address the potential harms of foundation models.

A growing ecosystem of large, open-source foundation models has reduced the labeled data and technical expertise necessary to apply machine learning to many new problems. Yet foundation models pose a clear dual-use risk, indiscriminately reducing the costs of building both harmful and beneficial machine learning systems. Policy tools such as restricted model access and export controls are the primary methods currently used to mitigate such dual-use risks. In this work, we review potential safe-release strategies and argue that both policymakers and AI researchers would benefit from fundamentally new technologies enabling more precise control over the downstream usage of open-source foundation models. We propose one such approach: the task blocking paradigm, in which foundation models are trained with an additional mechanism to impede adaptation to harmful tasks without sacrificing performance on desirable tasks. We call the resulting models self-destructing models, inspired by mechanisms that prevent adversaries from using tools for harmful purposes. We present an algorithm for training self-destructing models leveraging techniques from meta-learning and adversarial learning, which we call meta-learned adversarial censoring (MLAC). In a small-scale experiment, we show MLAC can largely prevent a BERT-style model from being re-purposed to perform gender identification without harming the model's ability to perform profession classification., Comment: v1 Presented at the First Workshop of Pre-training: Perspectives, Pitfalls, and Paths Forward (ICML, 2022) and New Frontiers in Adversarial Machine Learning Workshop (ICML, 2022); v2 Presented at the Sixth AAAI/ACM Conference on AI, Ethics, and Society (AIES, 2023)

Language models (LMs) are becoming the foundation for almost all major language technologies, but their capabilities, limitations, and risks are not well understood. We present Holistic Evaluation of Language Models (HELM) to improve the transparency of language models. First, we taxonomize the vast space of potential scenarios (i.e. use cases) and metrics (i.e. desiderata) that are of interest for LMs. Then we select a broad subset based on coverage and feasibility, noting what's missing or underrepresented (e.g. question answering for neglected English dialects, metrics for trustworthiness). Second, we adopt a multi-metric approach: We measure 7 metrics (accuracy, calibration, robustness, fairness, bias, toxicity, and efficiency) for each of 16 core scenarios when possible (87.5% of the time). This ensures metrics beyond accuracy don't fall to the wayside, and that trade-offs are clearly exposed. We also perform 7 targeted evaluations, based on 26 targeted scenarios, to analyze specific aspects (e.g. reasoning, disinformation). Third, we conduct a large-scale evaluation of 30 prominent language models (spanning open, limited-access, and closed models) on all 42 scenarios, 21 of which were not previously used in mainstream LM evaluation. Prior to HELM, models on average were evaluated on just 17.9% of the core HELM scenarios, with some prominent models not sharing a single scenario in common. We improve this to 96.0%: now all 30 models have been densely benchmarked on the same core scenarios and metrics under standardized conditions. Our evaluation surfaces 25 top-level findings. For full transparency, we release all raw model prompts and completions publicly for further analysis, as well as a general modular toolkit. We intend for HELM to be a living benchmark for the community, continuously updated with new scenarios, metrics, and models., Comment: Authored by the Center for Research on Foundation Models (CRFM) at the Stanford Institute for Human-Centered Artificial Intelligence (HAI). Project page: https://crfm.stanford.edu/helm/v1.0

Large language models (LLMs) have been shown to be able to perform new tasks based on a few demonstrations or natural language instructions. While these capabilities have led to widespread adoption, most LLMs are developed by resource-rich organizations and are frequently kept from the public. As a step towards democratizing this powerful technology, we present BLOOM, a 176B-parameter open-access language model designed and built thanks to a collaboration of hundreds of researchers. BLOOM is a decoder-only Transformer language model that was trained on the ROOTS corpus, a dataset comprising hundreds of sources in 46 natural and 13 programming languages (59 in total). We find that BLOOM achieves competitive performance on a wide variety of benchmarks, with stronger results after undergoing multitask prompted finetuning. To facilitate future research and applications using LLMs, we publicly release our models and code under the Responsible AI License.

In NLP, models are usually evaluated by reporting single-number performance scores on a number of readily available benchmarks, without much deeper analysis. Here, we argue that - especially given the well-known fact that benchmarks often contain biases, artefacts, and spurious correlations - deeper results analysis should become the de-facto standard when presenting new models or benchmarks. We present a tool that researchers can use to study properties of the dataset and the influence of those properties on their models' behaviour. Our Text Characterization Toolkit includes both an easy-to-use annotation tool, as well as off-the-shelf scripts that can be used for specific analyses. We also present use-cases from three different domains: we use the tool to predict what are difficult examples for given well-known trained models and identify (potentially harmful) biases and heuristics that are present in a dataset.

Entropy regularization is known to improve exploration in sequential decision-making problems. We show that this same mechanism can also lead to nearly unbiased and lower-variance estimates of the mean reward in the optimize-and-estimate structured bandit setting. Mean reward estimation (i.e., population estimation) tasks have recently been shown to be essential for public policy settings where legal constraints often require precise estimates of population metrics. We show that leveraging entropy and KL divergence can yield a better trade-off between reward and estimator variance than existing baselines, all while remaining nearly unbiased. These properties of entropy regularization illustrate an exciting potential for bridging the optimal exploration and estimation literatures.

One concern with the rise of large language models lies with their potential for significant harm, particularly from pretraining on biased, obscene, copyrighted, and private information. Emerging ethical approaches have attempted to filter pretraining material, but such approaches have been ad hoc and failed to take context into account. We offer an approach to filtering grounded in law, which has directly addressed the tradeoffs in filtering material. First, we gather and make available the Pile of Law, a 256GB (and growing) dataset of open-source English-language legal and administrative data, covering court opinions, contracts, administrative rules, and legislative records. Pretraining on the Pile of Law may help with legal tasks that have the promise to improve access to justice. Second, we distill the legal norms that governments have developed to constrain the inclusion of toxic or private content into actionable lessons for researchers and discuss how our dataset reflects these norms. Third, we show how the Pile of Law offers researchers the opportunity to learn such filtering rules directly from the data, providing an exciting new research direction in model-based processing., Comment: Presented at NeurIPS Datasets & Benchmarks (2022)

The recent emergence and adoption of Machine Learning technology, and specifically of Large Language Models, has drawn attention to the need for systematic and transparent management of language data. This work proposes an approach to global language data governance that attempts to organize data management amongst stakeholders, values, and rights. Our proposal is informed by prior work on distributed governance that accounts for human values and grounded by an international research collaboration that brings together researchers and practitioners from 60 countries. The framework we present is a multi-party international governance structure focused on language data, and incorporating technical and organizational tools needed to support its work., Comment: 32 pages: Full paper and Appendices; Association for Computing Machinery, New York, NY, USA, 2206-2222

We introduce a new setting, optimize-and-estimate structured bandits. Here, a policy must select a batch of arms, each characterized by its own context, that would allow it to both maximize reward and maintain an accurate (ideally unbiased) population estimate of the reward. This setting is inherent to many public and private sector applications and often requires handling delayed feedback, small data, and distribution shifts. We demonstrate its importance on real data from the United States Internal Revenue Service (IRS). The IRS performs yearly audits of the tax base. Two of its most important objectives are to identify suspected misreporting and to estimate the "tax gap" -- the global difference between the amount paid and true amount owed. Based on a unique collaboration with the IRS, we cast these two processes as a unified optimize-and-estimate structured bandit. We analyze optimize-and-estimate approaches to the IRS problem and propose a novel mechanism for unbiased population estimation that achieves rewards comparable to baseline approaches. This approach has the potential to improve audit efficacy, while maintaining policy-relevant estimates of the tax gap. This has important social consequences given that the current tax gap is estimated at nearly half a trillion dollars. We suggest that this problem setting is fertile ground for further research and we highlight its interesting challenges. The results of this and related research are currently being incorporated into the continual improvement of the IRS audit selection methods., Comment: Accepted to the Thirty-Seventh AAAI Conference On Artificial Intelligence (AAAI), 2023

Pupils with Special Educational Needs and Disability (SEND) have the greatest need for excellent teaching and are entitled to provision that supports achievement at, and enjoyment of, school. The attainment gap between pupils with SEND and their peers is twice as big as the gap between pupils eligible for free school meals and their peers. However, pupils with SEND are also more than twice as likely to be eligible for free school meals. This is why we've developed this guidance report. It offers five evidence-based recommendations to support pupils with SEND, providing a starting point for schools to review their current approach and practical ideas they can implement. [This guidance report was produced with the financial support of The Kusuma Trust UK. For the related evidence review "Special Educational Needs in Mainstream Schools: Evidence Review," see ED612232.]

We explore the promises and challenges of employing sequential decision-making algorithms -- such as bandits, reinforcement learning, and active learning -- in law and public policy. While such algorithms have well-characterized performance in the private sector (e.g., online advertising), the tendency to naively apply algorithms motivated by one domain, often online advertisements, can be called the "advertisement fallacy." Our main thesis is that law and public policy pose distinct methodological challenges that the machine learning community has not yet addressed. Machine learning will need to address these methodological problems to move "beyond ads." Public law, for instance, can pose multiple objectives, necessitate batched and delayed feedback, and require systems to learn rational, causal decision-making policies, each of which presents novel questions at the research frontier. We discuss a wide range of potential applications of sequential decision-making algorithms in regulation and governance, including public health, environmental protection, tax administration, occupational safety, and benefits adjudication. We use these examples to highlight research needed to render sequential decision making policy-compliant, adaptable, and effective in the public sector. We also note the potential risks of such deployments and describe how sequential decision systems can also facilitate the discovery of harms. We hope our work inspires more investigation of sequential decision making in law and public policy, which provide unique challenges for machine learning researchers with potential for significant social benefit., Comment: Version 1 presented at Causal Inference Challenges in Sequential Decision Making: Bridging Theory and Practice (2021), a NeurIPS 2021 Workshop; Version 2 presented at the 2nd ACM Symposium on Computer Science and Law (2022) (DOI: https://dl.acm.org/doi/10.1145/3511265.3550439)

AI is undergoing a paradigm shift with the rise of models (e.g., BERT, DALL-E, GPT-3) that are trained on broad data at scale and are adaptable to a wide range of downstream tasks. We call these models foundation models to underscore their critically central yet incomplete character. This report provides a thorough account of the opportunities and risks of foundation models, ranging from their capabilities (e.g., language, vision, robotics, reasoning, human interaction) and technical principles(e.g., model architectures, training procedures, data, systems, security, evaluation, theory) to their applications (e.g., law, healthcare, education) and societal impact (e.g., inequity, misuse, economic and environmental impact, legal and ethical considerations). Though foundation models are based on standard deep learning and transfer learning, their scale results in new emergent capabilities,and their effectiveness across so many tasks incentivizes homogenization. Homogenization provides powerful leverage but demands caution, as the defects of the foundation model are inherited by all the adapted models downstream. Despite the impending widespread deployment of foundation models, we currently lack a clear understanding of how they work, when they fail, and what they are even capable of due to their emergent properties. To tackle these questions, we believe much of the critical research on foundation models will require deep interdisciplinary collaboration commensurate with their fundamentally sociotechnical nature., Comment: Authored by the Center for Research on Foundation Models (CRFM) at the Stanford Institute for Human-Centered Artificial Intelligence (HAI). Report page with citation guidelines: https://crfm.stanford.edu/report.html

While self-supervised learning has made rapid advances in natural language processing, it remains unclear when researchers should engage in resource-intensive domain-specific pretraining (domain pretraining). The law, puzzlingly, has yielded few documented instances of substantial gains to domain pretraining in spite of the fact that legal language is widely seen to be unique. We hypothesize that these existing results stem from the fact that existing legal NLP tasks are too easy and fail to meet conditions for when domain pretraining can help. To address this, we first present CaseHOLD (Case Holdings On Legal Decisions), a new dataset comprised of over 53,000+ multiple choice questions to identify the relevant holding of a cited case. This dataset presents a fundamental task to lawyers and is both legally meaningful and difficult from an NLP perspective (F1 of 0.4 with a BiLSTM baseline). Second, we assess performance gains on CaseHOLD and existing legal NLP datasets. While a Transformer architecture (BERT) pretrained on a general corpus (Google Books and Wikipedia) improves performance, domain pretraining (using corpus of approximately 3.5M decisions across all courts in the U.S. that is larger than BERT's) with a custom legal vocabulary exhibits the most substantial performance gains with CaseHOLD (gain of 7.2% on F1, representing a 12% improvement on BERT) and consistent performance gains across two other legal tasks. Third, we show that domain pretraining may be warranted when the task exhibits sufficient similarity to the pretraining corpus: the level of performance increase in three legal tasks was directly tied to the domain specificity of the task. Our findings inform when researchers should engage resource-intensive pretraining and show that Transformer-based architectures, too, learn embeddings suggestive of distinct legal language., Comment: ICAIL 2021. Code & data available at https://github.com/reglab/casehold

How do we formalize the challenge of credit assignment in reinforcement learning? Common intuition would draw attention to reward sparsity as a key contributor to difficult credit assignment and traditional heuristics would look to temporal recency for the solution, calling upon the classic eligibility trace. We posit that it is not the sparsity of the reward itself that causes difficulty in credit assignment, but rather the \emph{information sparsity}. We propose to use information theory to define this notion, which we then use to characterize when credit assignment is an obstacle to efficient learning. With this perspective, we outline several information-theoretic mechanisms for measuring credit under a fixed behavior policy, highlighting the potential of information theory as a key tool towards provably-efficient credit assignment., Comment: Workshop on Biological and Artificial Reinforcement Learning (NeurIPS 2020)

Despite its importance to experimental design, statistical power (the probability that, given a real effect, an experiment will reject the null hypothesis) has largely been ignored by the NLP community. Underpowered experiments make it more difficult to discern the difference between statistical noise and meaningful model improvements, and increase the chances of exaggerated findings. By meta-analyzing a set of existing NLP papers and datasets, we characterize typical power for a variety of settings and conclude that underpowered experiments are common in the NLP literature. In particular, for several tasks in the popular GLUE benchmark, small test sets mean that most attempted comparisons to state of the art models will not be adequately powered. Similarly, based on reasonable assumptions, we find that the most typical experimental design for human rating studies will be underpowered to detect small model differences, of the sort that are frequently studied. For machine translation, we find that typical test sets of 2000 sentences have approximately 75% power to detect differences of 1 BLEU point. To improve the situation going forward, we give an overview of best practices for power analysis in NLP and release a series of notebooks to assist with future power analyses., Comment: To appear at EMNLP 2020

1) The annual cycle of atmospheric methane in southern high latitudes is extremely highly correlated with Antarctic sea ice extent. 2) The annual cycle of atmospheric methane in the Arctic is highly correlated with Antarctic or Arctic plus Antarctic sea ice extent. 3) We propose the global annual cycle of atmospheric methane is largely driven by Antarctic sea ice dynamics, with relatively stronger influence from other fluxes (probably the biota) in the Northern Hemisphere. 4) We propose degassing during sea ice freeze and temperature dependent solubility in the ocean dominate the annual methane cycle. 5) Results provide evidence that carbon cycle pathways, parameters and predictions must be reassessed., Comment: 27 pages including 2 appendices

This report documents ideas for improving the field of machine learning, which arose from discussions at the ML Retrospectives workshop at NeurIPS 2019. The goal of the report is to disseminate these ideas more broadly, and in turn encourage continuing discussion about how the field could improve along these axes. We focus on topics that were most discussed at the workshop: incentives for encouraging alternate forms of scholarship, re-structuring the review process, participation from academia and industry, and how we might better train computer scientists as scientists. Videos from the workshop can be accessed at https://slideslive.com/neurips/west-114-115-retrospectives-a-venue-for-selfreflection-in-ml-research

We investigate whether Jacobi preconditioning, accounting for the bootstrap term in temporal difference (TD) learning, can help boost performance of adaptive optimizers. Our method, TDprop, computes a per parameter learning rate based on the diagonal preconditioning of the TD update rule. We show how this can be used in both $n$-step returns and TD($\lambda$). Our theoretical findings demonstrate that including this additional preconditioning information is, surprisingly, comparable to normal semi-gradient TD if the optimal learning rate is found for both via a hyperparameter search. In Deep RL experiments using Expected SARSA, TDprop meets or exceeds the performance of Adam in all tested games under near-optimal learning rates, but a well-tuned SGD can yield similar improvements -- matching our theory. Our findings suggest that Jacobi preconditioning may improve upon typical adaptive optimization methods in Deep RL, but despite incorporating additional information from the TD bootstrap term, may not always be better than SGD., Comment: Presented at the Theoretical Foundations of Reinforcement Learning workshop at ICML 2020

With the recent wave of progress in artificial intelligence (AI) has come a growing awareness of the large-scale impacts of AI systems, and recognition that existing regulations and norms in industry and academia are insufficient to ensure responsible AI development. In order for AI developers to earn trust from system users, customers, civil society, governments, and other stakeholders that they are building AI responsibly, they will need to make verifiable claims to which they can be held accountable. Those outside of a given organization also need effective means of scrutinizing such claims. This report suggests various steps that different stakeholders can take to improve the verifiability of claims made about AI systems and their associated development processes, with a focus on providing evidence about the safety, security, fairness, and privacy protection of AI systems. We analyze ten mechanisms for this purpose--spanning institutions, software, and hardware--and make recommendations aimed at implementing, exploring, or improving those mechanisms.

Accurate reporting of energy and carbon usage is essential for understanding the potential climate impacts of machine learning research. We introduce a framework that makes this easier by providing a simple interface for tracking realtime energy consumption and carbon emissions, as well as generating standardized online appendices. Utilizing this framework, we create a leaderboard for energy efficient reinforcement learning algorithms to incentivize responsible research in this area as an example for other areas of machine learning. Finally, based on case studies using our framework, we propose strategies for mitigation of carbon emissions and reduction of energy consumption. By making accounting easier, we hope to further the sustainable development of machine learning experiments and spur more research into energy efficient algorithms., Comment: Published in JMLR: https://jmlr.org/papers/v21/20-312.html

In many finite horizon episodic reinforcement learning (RL) settings, it is desirable to optimize for the undiscounted return - in settings like Atari, for instance, the goal is to collect the most points while staying alive in the long run. Yet, it may be difficult (or even intractable) mathematically to learn with this target. As such, temporal discounting is often applied to optimize over a shorter effective planning horizon. This comes at the risk of potentially biasing the optimization target away from the undiscounted goal. In settings where this bias is unacceptable - where the system must optimize for longer horizons at higher discounts - the target of the value function approximator may increase in variance leading to difficulties in learning. We present an extension of temporal difference (TD) learning, which we call TD($\Delta$), that breaks down a value function into a series of components based on the differences between value functions with smaller discount factors. The separation of a longer horizon value function into these components has useful properties in scalability and performance. We discuss these properties and show theoretic and empirical improvements over standard TD learning in certain settings., Comment: Full version accepted to ICML 2019. Extended abstract also to be presented at RLDM 2019