ShiQ: Bringing back Bellman to LLMs

• URL: http://arxiv.org/abs/2505.11081v1
• Date: Fri, 16 May 2025 10:12:11 GMT
• Title: ShiQ: Bringing back Bellman to LLMs
• Authors: Pierre Clavier, Nathan Grinsztajn, Raphael Avalos, Yannis Flet-Berliac, Irem Ergun, Omar D. Domingues, Eugene Tarassov, Olivier Pietquin, Pierre H. Richemond, Florian Strub, Matthieu Geist,
• Abstract summary: We build a practical algorithm, ShiQ for Shifted-Q, that supports off-policy, token-wise learning while remaining simple to implement.<n>We evaluate ShiQ on both synthetic data and real-world benchmarks, e.g., UltraFeedback and BFCL-V3, demonstrating its effectiveness in both single-turn and multi-turn LLM settings.
• Score: 37.70964838115103
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: The fine-tuning of pre-trained large language models (LLMs) using reinforcement learning (RL) is generally formulated as direct policy optimization. This approach was naturally favored as it efficiently improves a pretrained LLM, seen as an initial policy. Another RL paradigm, Q-learning methods, has received far less attention in the LLM community while demonstrating major success in various non-LLM RL tasks. In particular, Q-learning effectiveness comes from its sample efficiency and ability to learn offline, which is particularly valuable given the high computational cost of sampling with LLMs. However, naively applying a Q-learning-style update to the model's logits is ineffective due to the specificity of LLMs. Our core contribution is to derive theoretically grounded loss functions from Bellman equations to adapt Q-learning methods to LLMs. To do so, we carefully adapt insights from the RL literature to account for LLM-specific characteristics, ensuring that the logits become reliable Q-value estimates. We then use this loss to build a practical algorithm, ShiQ for Shifted-Q, that supports off-policy, token-wise learning while remaining simple to implement. Finally, we evaluate ShiQ on both synthetic data and real-world benchmarks, e.g., UltraFeedback and BFCL-V3, demonstrating its effectiveness in both single-turn and multi-turn LLM settings

Related papers

• A Little Help Goes a Long Way: Efficient LLM Training by Leveraging Small LMs [74.35290684163718]
  A primary challenge in large language model (LLM) development is their onerous pre-training cost. This paper explores a promising paradigm to improve LLM pre-training efficiency and quality by leveraging a small language model (SLM)
  arXiv  Detail & Related papers  (2024-10-24T14:31:52Z)
• VerifierQ: Enhancing LLM Test Time Compute with Q-Learning-based Verifiers [7.7705926659081275]
  VerifierQ is a novel approach that integrates Offline Q-learning into verifier models. We address three key challenges in applying Q-learning to LLMs. Our method enables parallel Q-value computation and improving training efficiency.
  arXiv  Detail & Related papers  (2024-10-10T15:43:55Z)
• Towards Self-Improvement of LLMs via MCTS: Leveraging Stepwise Knowledge with Curriculum Preference Learning [70.16816087320585]
  Monte Carlo Tree Search (MCTS) has emerged as a powerful technique for enhancing the reasoning capabilities of LLMs. Existing distillation methods underutilize the rich trajectory information generated by MCTS. We propose AlphaLLM-CPL, a novel pairwise training framework that enables LLMs to self-improve through MCTS behavior distillation.
  arXiv  Detail & Related papers  (2024-10-09T03:20:02Z)
• VinePPO: Refining Credit Assignment in RL Training of LLMs [66.80143024475635]
  We propose VinePPO, a straightforward approach that leverages the flexibility of language environments to compute unbiased Monte Carlo-based estimates.<n>Our method consistently outperforms PPO and other baselines across MATH and GSM8K datasets in less wall-clock time.
  arXiv  Detail & Related papers  (2024-10-02T15:49:30Z)
• Q-Sparse: All Large Language Models can be Fully Sparsely-Activated [93.45300714803429]
  We introduce Q-Sparse, a simple yet effective approach to training sparsely-activated large language models (LLMs) Q-Sparse enables full sparsity of activations in LLMs which can bring significant efficiency gains in inference. We also introduce Block Q-Sparse for batch training and inference.
  arXiv  Detail & Related papers  (2024-07-15T17:59:29Z)
• Q-Adapter: Customizing Pre-trained LLMs to New Preferences with Forgetting Mitigation [35.014856057848036]
  We consider customizing pre-trained Large Language Models (LLMs) with new human preferences.<n>We name our method Q-Adapter as it introduces an adapter module to approximate the residual Q-function for customizing the new preference.<n> Experiments based on the Llama-3.1 model on the DSP dataset and HH-RLHF dataset illustrate the superior effectiveness of Q-Adapter.
  arXiv  Detail & Related papers  (2024-07-04T11:42:36Z)
• Q*: Improving Multi-step Reasoning for LLMs with Deliberative Planning [53.6472920229013]
  Large Language Models (LLMs) have demonstrated impressive capability in many natural language tasks. LLMs are prone to produce errors, hallucinations and inconsistent statements when performing multi-step reasoning. We introduce Q*, a framework for guiding LLMs decoding process with deliberative planning.
  arXiv  Detail & Related papers  (2024-06-20T13:08:09Z)
• How Can LLM Guide RL? A Value-Based Approach [68.55316627400683]
  Reinforcement learning (RL) has become the de facto standard practice for sequential decision-making problems by improving future acting policies with feedback. Recent developments in large language models (LLMs) have showcased impressive capabilities in language understanding and generation, yet they fall short in exploration and self-improvement capabilities. We develop an algorithm named LINVIT that incorporates LLM guidance as a regularization factor in value-based RL, leading to significant reductions in the amount of data needed for learning.
  arXiv  Detail & Related papers  (2024-02-25T20:07:13Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.