Pith. sign in

REVIEW 6 cited by

LLMs are Greedy Agents: Effects of RL Fine-tuning on Decision-Making Abilities

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2504.16078 v1 pith:NVLQFEDO submitted 2025-04-22 cs.LG cs.AI

classification cs.LGcs.AI
keywords llmsdecision-makingfine-tuningexplorationknowing-doingabilitiesagentsbandits
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
abstract

The success of Large Language Models (LLMs) has sparked interest in various agentic applications. A key hypothesis is that LLMs, leveraging common sense and Chain-of-Thought (CoT) reasoning, can effectively explore and efficiently solve complex domains. However, LLM agents have been found to suffer from sub-optimal exploration and the knowing-doing gap, the inability to effectively act on knowledge present in the model. In this work, we systematically study why LLMs perform sub-optimally in decision-making scenarios. In particular, we closely examine three prevalent failure modes: greediness, frequency bias, and the knowing-doing gap. We propose mitigation of these shortcomings by fine-tuning via Reinforcement Learning (RL) on self-generated CoT rationales. Our experiments across multi-armed bandits, contextual bandits, and Tic-tac-toe, demonstrate that RL fine-tuning enhances the decision-making abilities of LLMs by increasing exploration and narrowing the knowing-doing gap. Finally, we study both classic exploration mechanisms, such as $\epsilon$-greedy, and LLM-specific approaches, such as self-correction and self-consistency, to enable more effective fine-tuning of LLMs for decision-making.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 6 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. STOCKTAKE: Measuring the Gap Between Perception and Action in LLM Agents with a Fair Oracle

    cs.AI 2026-07 conditional novelty 7.0 of 10

    Frontier LLM agents detect hidden supply-chain stress almost equally well (84–88% of episodes) but vary from skill 0.62 to −0.23, with two of four models acting worse than ignoring symptoms.

  2. Imaging-101: Benchmarking LLM Coding Agents on Scientific Computational Imaging

    cs.AI 2026-07 conditional novelty 7.0 of 10

    A 57-task, expert-verified benchmark shows frontier LLM coding agents systematically fail on physical conventions, inverse-solver choice, and end-to-end imaging pipelines.

  3. Formalizing Learning from Language Feedback with Provable Guarantees

    cs.LG 2025-06 conditional novelty 7.0 of 10

    Introduces a formal framework for learning from language feedback, a transfer eluder dimension complexity measure, and HELiX, a no-regret algorithm whose regret scales with this dimension.

  4. BioBlue: Systematic runaway-optimiser-like LLM failure modes on biologically and economically aligned AI safety benchmarks for LLMs with simplified observation format

    cs.CY 2025-09 conditional novelty 6.0 of 10

    LLMs in long-horizon multi-objective simulations show a recurrent drift from balanced, target-following behavior to single-objective, unbounded maximization, despite initial competence.

  5. AIGB-R1: Self-Evolving Generative Auto-Bidding via Hierarchical Planner-Executor Optimization

    cs.LG 2026-07 conditional novelty 5.0 of 10

    AIGB-R1 couples an LLM strategy planner with a prompt-conditioned Decision Transformer executor and trains both end-to-end with a decoupled-GRPO self-evolving loop, reporting the highest Score on the AuctionNet benchmark.

  6. Reward Modeling for Reinforcement Learning-Based LLM Reasoning: Design, Challenges, and Evaluation

    cs.LG 2026-02 conditional novelty 4.0 of 10

    A taxonomy-driven survey arguing that reward design is the central mechanism shaping reliable LLM reasoning, with maps of reward paradigms, reward-hacking failure modes, and benchmark pitfalls.

Pith tools