Search

The reward model has become increasingly important in alignment, assessment, and data construction for large language models (LLMs). Most existing researchers focus on enhancing reward models through data improvements, following the conventional training framework for reward models that directly optimizes the predicted rewards. In this paper, we propose a hybrid alignment framework HaF-RM for reward model training by introducing an additional constraint on token-level policy probabilities in addition to the reward score. It can simultaneously supervise the internal preference model at the token level and optimize the mapping layer of the reward model at the sequence level. Theoretical justifications and experiment results on five datasets show the validity and effectiveness of our proposed hybrid framework for training a high-quality reward model. By decoupling the reward modeling procedure and incorporating hybrid supervision, our HaF-RM framework offers a principled and effective approach to enhancing the performance and alignment of reward models, a critical component in the responsible development of powerful language models. We release our code at https://haf-rm.github.io.

We introduce ALaRM, the first framework modeling hierarchical rewards in reinforcement learning from human feedback (RLHF), which is designed to enhance the alignment of large language models (LLMs) with human preferences. The framework addresses the limitations of current alignment approaches, which often struggle with the inconsistency and sparsity of human supervision signals, by integrating holistic rewards with aspect-specific rewards. This integration enables more precise and consistent guidance of language models towards desired outcomes, particularly in complex and open text generation tasks. By employing a methodology that filters and combines multiple rewards based on their consistency, the framework provides a reliable mechanism for improving model alignment. We validate our approach through applications in long-form question answering and machine translation tasks, employing gpt-3.5-turbo for pairwise comparisons, and demonstrate improvements over existing baselines. Our work underscores the effectiveness of hierarchical rewards modeling in refining LLM training processes for better human preference alignment. We release our code at https://ALaRM-fdu.github.io., Comment: 15 pages, 6 figures

We propose DISC-LawLLM, an intelligent legal system utilizing large language models (LLMs) to provide a wide range of legal services. We adopt legal syllogism prompting strategies to construct supervised fine-tuning datasets in the Chinese Judicial domain and fine-tune LLMs with legal reasoning capability. We augment LLMs with a retrieval module to enhance models' ability to access and utilize external legal knowledge. A comprehensive legal benchmark, DISC-Law-Eval, is presented to evaluate intelligent legal systems from both objective and subjective dimensions. Quantitative and qualitative results on DISC-Law-Eval demonstrate the effectiveness of our system in serving various users across diverse legal scenarios. The detailed resources are available at https://github.com/FudanDISC/DISC-LawLLM.

Transport of fast electron in overdense plasmas is of key importance in high energy density physics. However, it is challenging to diagnose the fast electron transport in experiments. In this article, we study coherent transition radiation (CTR) generated by fast electrons on the back surface of the target by using 2D and 3D first-principle particle-in-cell (PIC) simulations. In our simulations, aluminium target of 2.7 g/cc is simulated in two different situations by using a newly developed high order implicit PIC code. Comparing realistic simulations containing collision and ionization effects, artificial simulations without taking collision and ionization effects into account significantly underestimate the energy loss of electron beam when transporting in the target, which fail to describe the complete characteristics of CTR produced by electron beam on the back surface of the target. Realistic simulations indicate the diameter of CTR increases when the thickness of the target is increased. This is attributed to synergetic energy losses of high flux fast electrons due to Ohm heatings and colliding drags, which appear quite significant even when the thickness of the solid target only differs by micrometers. Especially, when the diagnosing position is fixed, we find that the intensity distribution of the CTR is also a function of time, with the diameter increased with time. As the diameter of CTR is related to the speed of electrons passing through the back surface of the target, our finding may be used as a new tool to diagnose the electron energy spectra near the surface of solid density plasmas., Comment: accepted by New Journal of Physics

Suffering from limited singing voice corpus, existing singing voice synthesis (SVS) methods that build encoder-decoder neural networks to directly generate spectrogram could lead to out-of-tune issues during the inference phase. To attenuate these issues, this paper presents a novel acoustic model with independent pitch encoder and phoneme encoder, which disentangles the phoneme and pitch information from music score to fully utilize the corpus. Specifically, according to equal temperament theory, the pitch encoder is constrained by a pitch metric loss that maps distances between adjacent input pitches into corresponding frequency multiples between the encoder outputs. For the phoneme encoder, based on the analysis that same phonemes corresponding to varying pitches can produce similar pronunciations, this encoder is followed by an adversarially trained pitch classifier to enforce the identical phonemes with different pitches mapping into the same phoneme feature space. By these means, the sparse phonemes and pitches in original input spaces can be transformed into more compact feature spaces respectively, where the same elements cluster closely and cooperate mutually to enhance synthesis quality. Then, the outputs of the two encoders are summed together to pass through the following decoder in the acoustic model. Experimental results indicate that the proposed approaches can characterize intrinsic structure between pitch inputs to obtain better pitch synthesis accuracy and achieve superior singing synthesis performance against the advanced baseline system., Comment: 5 pages, 3 figures