When Greedy Wins: Emergent Exploitation Bias in Meta-Bandit LLM Training

• URL: http://arxiv.org/abs/2509.24923v1
• Date: Mon, 29 Sep 2025 15:25:42 GMT
• Title: When Greedy Wins: Emergent Exploitation Bias in Meta-Bandit LLM Training
• Authors: Sanxing Chen, Xiaoyin Chen, Yukun Huang, Roy Xie, Bhuwan Dhingra,
• Abstract summary: Large Language Models (LLMs) often explore suboptimally in sequential decision-making.<n>Recent work has sought to enhance this capability via supervised fine-tuning (SFT) or reinforcement learning (RL), improving regret on the classic multi-armed bandit task.<n>We investigate both paradigms by training LLMs with SFT on expert trajectories and RL with a range of tailored reward signals.<n>The resulting agents outperform pre-trained models and achieve performance comparable to Upper Confidence Bound (UCB) and Thompson Sampling.
• Score: 26.66184262287797
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: While Large Language Models (LLMs) hold promise to become autonomous agents, they often explore suboptimally in sequential decision-making. Recent work has sought to enhance this capability via supervised fine-tuning (SFT) or reinforcement learning (RL), improving regret on the classic multi-armed bandit task. However, it remains unclear how these learning methods shape exploration strategies and how well they generalize. We investigate both paradigms by training LLMs with SFT on expert trajectories and RL with a range of tailored reward signals including a strategic, regret-shaped reward to reduce variance, and an algorithmic reward that enables oracle imitation. The resulting agents outperform pre-trained models and achieve performance comparable to Upper Confidence Bound (UCB) and Thompson Sampling, with robust generalization to 6x longer horizons and across bandit families. Behavioral analysis reveals that gains often stem from more sophisticated but greedier exploitation: RL/SFT agents are more prone to early catastrophic failure than pre-trained models, prematurely abandoning exploration. Furthermore, agents trained to imitate UCB learn to outperform their teacher by adopting more exploitative variants. Our findings clarify when each training paradigm is preferable and advocate tailored reward design and evaluation beyond average regret to promote robust exploratory behavior.

Related papers

• When Actions Teach You to Think: Reasoning-Action Synergy via Reinforcement Learning in Conversational Agents [2.689316553293938]
  Supervised fine-tuning (SFT) has emerged as one of the most effective ways to improve the performance of large language models (LLMs) in downstream tasks.<n>We propose a pipeline in which LLMs generate reasoning steps that guide both the invocation of tools and the final answer generation for conversational agents.
  arXiv  Detail & Related papers  (2025-12-12T04:44:40Z)
• Multimodal Reinforcement Learning with Agentic Verifier for AI Agents [131.46008226323423]
  Argos is a principled multimodal reward agent to train reasoning models for agentic tasks.<n>By leveraging our agentic verifier across both SFT data and RL training, our model achieves state-of-the-art results.
  arXiv  Detail & Related papers  (2025-12-03T04:42:47Z)
• Explore Data Left Behind in Reinforcement Learning for Reasoning Language Models [61.78513830395669]
  Reinforcement Learning with Verifiable Rewards (RLVR) has emerged as an effective approach for improving the reasoning abilities of large language models (LLMs)<n>As models train longer and scale larger, more training prompts become residual prompts, those with zero variance rewards that provide no training signal.<n>We propose the Explore Residual Prompts in Policy Optimization framework, which encourages exploration on residual prompts and reactivates their training signals.
  arXiv  Detail & Related papers  (2025-11-06T20:40:27Z)
• MENTOR: A Reinforcement Learning Framework for Enabling Tool Use in Small Models via Teacher-Optimized Rewards [8.645370827540996]
  Distilling the tool-using capabilities of large language models (LLMs) into smaller, more efficient small language models (SLMs) is a key challenge for their practical application.<n>The predominant approach, supervised fine-tuning (SFT), suffers from poor generalization as it trains models to imitate a static set of teacher trajectories rather than learn a robust methodology.<n>We propose MENTOR, a framework that combines reinforcement learning with teacher-guided distillation.
  arXiv  Detail & Related papers  (2025-10-21T08:03:14Z)
• Agentic Reinforcement Learning with Implicit Step Rewards [92.26560379363492]
  Large language models (LLMs) are increasingly developed as autonomous agents using reinforcement learning (agentic RL)<n>We introduce implicit step rewards for agentic RL (iStar), a general credit-assignment strategy that integrates seamlessly with standard RL algorithms.<n>We evaluate our method on three challenging agent benchmarks, including WebShop and VisualSokoban, as well as open-ended social interactions with unverifiable rewards in SOTOPIA.
  arXiv  Detail & Related papers  (2025-09-23T16:15:42Z)
• RLVMR: Reinforcement Learning with Verifiable Meta-Reasoning Rewards for Robust Long-Horizon Agents [43.806220882212386]
  RLVMR integrates dense, process-level supervision into end-to-end RL by rewarding verifiable, meta-reasoning behaviors.<n>On the challenging ALFWorld and ScienceWorld benchmarks, RLVMR achieves new state-of-the-art results.
  arXiv  Detail & Related papers  (2025-07-30T17:00:48Z)
• Beyond Markovian: Reflective Exploration via Bayes-Adaptive RL for LLM Reasoning [55.36978389831446]
  We recast reflective exploration within the Bayes-Adaptive RL framework.<n>Our resulting algorithm, BARL, instructs the LLM to stitch and switch strategies based on observed outcomes.
  arXiv  Detail & Related papers  (2025-05-26T22:51:00Z)
• On the Diminishing Returns of Complex Robust RAG Training in the Era of Powerful LLMs [85.688901949146]
  We investigate the question: does the benefit of complex robust training methods diminish as language models become more powerful?<n>Our analysis reveals a consistent trend: emphthe marginal robustness benefit of sophisticated training strategies decreases substantially as model capacity increases.<n>Further investigation demonstrates that stronger models naturally exhibit better confidence calibration, cross-dataset generalization capability, and more effective attention patterns, even under simple training regimes.
  arXiv  Detail & Related papers  (2025-02-17T03:34:31Z)
• From Novice to Expert: LLM Agent Policy Optimization via Step-wise Reinforcement Learning [62.54484062185869]
  We introduce StepAgent, which utilizes step-wise reward to optimize the agent's reinforcement learning process.<n>We propose implicit-reward and inverse reinforcement learning techniques to facilitate agent reflection and policy adjustment.
  arXiv  Detail & Related papers  (2024-11-06T10:35:11Z)
• Learning to Optimize for Reinforcement Learning [58.01132862590378]
  Reinforcement learning (RL) is essentially different from supervised learning, and in practice, these learneds do not work well even in simple RL tasks. Agent-gradient distribution is non-independent and identically distributed, leading to inefficient meta-training. We show that, although only trained in toy tasks, our learned can generalize unseen complex tasks in Brax.
  arXiv  Detail & Related papers  (2023-02-03T00:11:02Z)

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.