NOVER: Incentive Training for Language Models via Verifier-Free Reinforcement Learning

Recent advances such as DeepSeek R1-Zero highlight the effectiveness of incentive training, a reinforcement learning paradigm that computes rewards solely based on the final answer part of a language model's output, thereby encouraging the generation of intermediate reasoning steps. However, these methods fundamentally rely on external verifiers, which limits their applicability to domains like mathematics and coding where such verifiers are readily available. Although reward models can serve as verifiers, they require high-quality annotated data and are costly to train. In this work, we propose NOVER, NO-VERifier Reinforcement Learning, a general reinforcement learning framework that requires only standard supervised fine-tuning data with no need for an external verifier. NOVER enables incentive training across a wide range of text-to-text tasks and outperforms the model of the same size distilled from large reasoning models such as DeepSeek R1 671B by 7.7 percent. Moreover, the flexibility of NOVER enables new possibilities for optimizing large language models, such as inverse incentive training.

Automatic Dataset Generation for Knowledge Intensive Question Answering Tasks

A question-answering (QA) system is to search suitable answers within a knowledge base. Current QA systems struggle with queries requiring complex reasoning or real-time knowledge integration. They are often supplemented with retrieval techniques on a data source such as Retrieval-Augmented Generation (RAG). However, RAG continues to face challenges in handling complex reasoning and logical connections between multiple sources of information. A novel approach for enhancing Large Language Models (LLMs) in knowledge-intensive QA tasks is presented through the automated generation of context-based QA pairs. This methodology leverages LLMs to create fine-tuning data, reducing reliance on human labelling and improving model comprehension and reasoning capabilities. The proposed system includes an automated QA generator and a model fine-tuner, evaluated using perplexity, ROUGE, BLEU, and BERTScore. Comprehensive experiments demonstrate improvements in logical coherence and factual accuracy, with implications for developing adaptable Artificial Intelligence (AI) systems. Mistral-7b-v0.3 outperforms Llama-3-8b with BERT F1, BLEU, and ROUGE scores 0.858, 0.172, and 0.260 of for the LLM generated QA pairs compared to scores of 0.836, 0.083, and 0.139 for the human annotated QA pairs.

Post-Incorporating Code Structural Knowledge into LLMs via In-Context Learning for Code Translation

Code translation migrates codebases across programming languages. Recently, large language models (LLMs) have achieved significant advancements in software mining. However, handling the syntactic structure of source code remains a challenge. Classic syntax-aware methods depend on intricate model architectures and loss functions, rendering their integration into LLM training resource-intensive. This paper employs in-context learning (ICL), which directly integrates task exemplars into the input context, to post-incorporate code structural knowledge into pre-trained LLMs. We revisit exemplar selection in ICL from an information-theoretic perspective, proposing that list-wise selection based on information coverage is more precise and general objective than traditional methods based on combining similarity and diversity. To address the challenges of quantifying information coverage, we introduce a surrogate measure, Coverage of Abstract Syntax Tree (CAST). Furthermore, we formulate the NP-hard CAST maximization for exemplar selection and prove that it is a standard submodular maximization problem. Therefore, we propose a greedy algorithm for CAST submodular maximization, which theoretically guarantees a (1-1/e)-approximate solution in polynomial time complexity. Our method is the first training-free and model-agnostic approach to post-incorporate code structural knowledge into existing LLMs at test time. Experimental results show that our method significantly improves LLMs performance and reveals two meaningful insights: 1) Code structural knowledge can be effectively post-incorporated into pre-trained LLMs during inference, despite being overlooked during training; 2) Scaling up model size or training data does not lead to the emergence of code structural knowledge, underscoring the necessity of explicitly considering code syntactic structure.

SocialJax: An Evaluation Suite for Multi-agent Reinforcement Learning in Sequential Social Dilemmas

Social dilemmas pose a significant challenge in the field of multi-agent reinforcement learning (MARL). Melting Pot is an extensive framework designed to evaluate social dilemma environments, providing an evaluation protocol that measures generalization to new social partners across various test scenarios. However, running reinforcement learning algorithms in the official Melting Pot environments demands substantial computational resources. In this paper, we introduce SocialJax, a suite of sequential social dilemma environments implemented in JAX. JAX is a high-performance numerical computing library for Python that enables significant improvements in the operational efficiency of SocialJax on GPUs and TPUs. Our experiments demonstrate that the training pipeline of SocialJax achieves a 50\texttimes{} speedup in real-time performance compared to Melting Pot's RLlib baselines. Additionally, we validate the effectiveness of baseline algorithms within the SocialJax environments. Finally, we use Schelling diagrams to verify the social dilemma properties of these environments, ensuring they accurately capture the dynamics of social dilemmas.

GRU: Mitigating the Trade-off between Unlearning and Retention for Large Language Models

Large language model (LLM) unlearning has demonstrated its essential role in removing privacy and copyright-related responses, crucial for their legal and safe applications. However, the pursuit of complete unlearning often comes with substantial costs due to its compromises in their general functionality, leading to a notorious trade-off between unlearning and retention. In examining the update process for unlearning dynamically, we find gradients hold essential information for revealing this trade-off. In particular, we look at the varying relationship between retention performance and directional disparities between gradients during unlearning. It motivates the sculpting of an update mechanism derived from gradients from two sources, i.e., harmful for retention and useful for unlearning. Accordingly, we propose Gradient Rectified Unlearning (GRU), an enhanced unlearning framework controlling the updating gradients in a geometry-focused and optimization-driven manner such that their side impacts on other, unrelated responses can be minimized. Specifically, GRU derives a closed-form solution to project the unlearning gradient onto the orthogonal space of that gradient harmful for retention, ensuring minimal deviation from its original direction under the condition that overall performance is retained. Comprehensive experiments are conducted to demonstrate that GRU, as a general framework, is straightforward to implement and efficiently enhances a range of baseline methods through its adaptable and compatible characteristics. Additionally, experimental results show its broad effectiveness across a diverse set of benchmarks for LLM unlearning.

M3HF: Multi-agent Reinforcement Learning from Multi-phase Human Feedback of Mixed Quality

Designing effective reward functions in multi-agent reinforcement learning (MARL) is a significant challenge, often leading to suboptimal or misaligned behaviors in complex, coordinated environments. We introduce Multi-agent Reinforcement Learning from Multi-phase Human Feedback of Mixed Quality (M3HF), a novel framework that integrates multi-phase human feedback of mixed quality into the MARL training process. By involving humans with diverse expertise levels to provide iterative guidance, M3HF leverages both expert and non-expert feedback to continuously refine agents' policies. During training, we strategically pause agent learning for human evaluation, parse feedback using large language models to assign it appropriately and update reward functions through predefined templates and adaptive weight by using weight decay and performance-based adjustments. Our approach enables the integration of nuanced human insights across various levels of quality, enhancing the interpretability and robustness of multi-agent cooperation. Empirical results in challenging environments demonstrate that M3HF significantly outperforms state-of-the-art methods, effectively addressing the complexities of reward design in MARL and enabling broader human participation in the training process.

ATLaS: Agent Tuning via Learning Critical Steps

Large Language Model (LLM) agents have demonstrated remarkable generalization capabilities across multi-domain tasks. Existing agent tuning approaches typically employ supervised finetuning on entire expert trajectories. However, behavior-cloning of full trajectories can introduce expert bias and weaken generalization to states not covered by the expert data. Additionally, critical steps, such as planning, complex reasoning for intermediate subtasks, and strategic decision-making, are essential to success in agent tasks, so learning these steps is the key to improving LLM agents. For more effective and efficient agent tuning, we propose ATLaS that identifies the critical steps in expert trajectories and finetunes LLMs solely on these steps with reduced costs. By steering the training's focus to a few critical steps, our method mitigates the risk of overfitting entire trajectories and promotes generalization across different environments and tasks. In extensive experiments, an LLM finetuned on only 30% critical steps selected by ATLaS outperforms the LLM finetuned on all steps and recent open-source LLM agents. ATLaS maintains and improves base LLM skills as generalist agents interacting with diverse environments.

$\text{M}^3\text{HF}$: Multi-agent Reinforcement Learning from Multi-phase Human Feedback of Mixed Quality

Designing effective reward functions in multi-agent reinforcement learning (MARL) is a significant challenge, often leading to suboptimal or misaligned behaviors in complex, coordinated environments. We introduce Multi-agent Reinforcement Learning from Multi-phase Human Feedback of Mixed Quality ($\text{M}^3\text{HF}$), a novel framework that integrates multi-phase human feedback of mixed quality into the MARL training process. By involving humans with diverse expertise levels to provide iterative guidance, $\text{M}^3\text{HF}$ leverages both expert and non-expert feedback to continuously refine agents' policies. During training, we strategically pause agent learning for human evaluation, parse feedback using large language models to assign it appropriately and update reward functions through predefined templates and adaptive weight by using weight decay and performance-based adjustments. Our approach enables the integration of nuanced human insights across various levels of quality, enhancing the interpretability and robustness of multi-agent cooperation. Empirical results in challenging environments demonstrate that $\text{M}^3\text{HF}$ significantly outperforms state-of-the-art methods, effectively addressing the complexities of reward design in MARL and enabling broader human participation in the training process.

CODI: Compressing Chain-of-Thought into Continuous Space via Self-Distillation

Chain-of-Thought (CoT) enhances Large Language Models (LLMs) by enabling step-by-step reasoning in natural language. However, the language space may be suboptimal for reasoning. While implicit CoT methods attempt to enable reasoning without explicit CoT tokens, they have consistently lagged behind explicit CoT method in task performance. We propose CODI (Continuous Chain-of-Thought via Self-Distillation), a novel framework that distills CoT into a continuous space, where a shared model acts as both teacher and student, jointly learning explicit and implicit CoT while aligning their hidden activation on the token generating the final answer. CODI is the first implicit CoT method to match explicit CoT's performance on GSM8k while achieving 3.1x compression, surpassing the previous state-of-the-art by 28.2% in accuracy. Furthermore, CODI demonstrates scalability, robustness, and generalizability to more complex CoT datasets. Additionally, CODI retains interpretability by decoding its continuous thoughts, making its reasoning process transparent. Our findings establish implicit CoT as not only a more efficient but a powerful alternative to explicit CoT.

Distill Not Only Data but Also Rewards: Can Smaller Language Models Surpass Larger Ones?

Distilling large language models (LLMs) typically involves transferring the teacher model's responses through supervised fine-tuning (SFT). However, this approach neglects the potential to distill both data (output content) and reward signals (quality evaluations). Extracting reliable reward signals directly from teacher models is challenging, as LLMs are optimized for generation rather than evaluation, often resulting in biased or inconsistent assessments. To address this limitation, we propose a novel distillation pipeline that transfers both responses and rewards. Our method generates pseudo-rewards through a self-supervised mechanism that leverages the inherent structure of both teacher and student responses, enabling reward learning without explicit external evaluation. The reward model subsequently guides reinforcement learning (RL), allowing iterative refinement of the student model after an SFT warm-up phase. Experiments on GSM8K and MMLU-PRO demonstrate that our method consistently outperforms traditional SFT-based approaches, enabling student models to surpass the performance of their teachers. This work highlights the potential for scalable, efficient distillation through structured self-supervised reward learning, reducing dependence on external reward supervision.