Large Language Models Reasoning Abilities Under Non-Ideal Conditions After RL-Fine-Tuning

• URL: http://arxiv.org/abs/2508.04848v1
• Date: Wed, 06 Aug 2025 19:51:29 GMT
• Title: Large Language Models Reasoning Abilities Under Non-Ideal Conditions After RL-Fine-Tuning
• Authors: Chang Tian, Matthew B. Blaschko, Mingzhe Xing, Xiuxing Li, Yinliang Yue, Marie-Francine Moens,
• Abstract summary: We introduce a new research direction guided by brain-science findings that human reasoning remains reliable under imperfect inputs.<n>We fine-tune three large language models (LLMs) and a state-of-the-art large vision-language model (LVLM) usingReinforcement learning (RL) with a representative policy-gradient algorithm.<n>Our results reveal that while RL fine-tuning improves baseline reasoning under idealized settings, performance declines significantly across all three non-ideal scenarios.
• Score: 33.27410930782468
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Reinforcement learning (RL) has become a key technique for enhancing the reasoning abilities of large language models (LLMs), with policy-gradient algorithms dominating the post-training stage because of their efficiency and effectiveness. However, most existing benchmarks evaluate large-language-model reasoning under idealized settings, overlooking performance in realistic, non-ideal scenarios. We identify three representative non-ideal scenarios with practical relevance: summary inference, fine-grained noise suppression, and contextual filtering. We introduce a new research direction guided by brain-science findings that human reasoning remains reliable under imperfect inputs. We formally define and evaluate these challenging scenarios. We fine-tune three LLMs and a state-of-the-art large vision-language model (LVLM) using RL with a representative policy-gradient algorithm and then test their performance on eight public datasets. Our results reveal that while RL fine-tuning improves baseline reasoning under idealized settings, performance declines significantly across all three non-ideal scenarios, exposing critical limitations in advanced reasoning capabilities. Although we propose a scenario-specific remediation method, our results suggest current methods leave these reasoning deficits largely unresolved. This work highlights that the reasoning abilities of large models are often overstated and underscores the importance of evaluating models under non-ideal scenarios. The code and data will be released at XXXX.

Related papers

• Learning to Reason as Action Abstractions with Scalable Mid-Training RL [55.24192942739207]
  An effective mid-training phase should identify a compact set of useful actions and enable fast selection.<n>We propose Reasoning as Action Abstractions (RA3), a scalable mid-training algorithm.
  arXiv  Detail & Related papers  (2025-09-30T05:34:20Z)
• Dissecting Long-Chain-of-Thought Reasoning Models: An Empirical Study [91.78803511141975]
  This work focuses on the roles of positive and negative samples in scaling reinforcement learning.<n>We identify substantial data inefficiency in group relative policy optimization, where over half of the samples yield zero advantage.<n>We investigate unstable performance across various reasoning models and benchmarks, attributing instability to uncertain problems with ambiguous outcomes.
  arXiv  Detail & Related papers  (2025-06-05T11:47:10Z)
• Beyond Accuracy: Dissecting Mathematical Reasoning for LLMs Under Reinforcement Learning [82.43575191712726]
  We introduce a fine-grained analytic framework to dissect the impact ofReinforcement learning on reasoning.<n>Our framework specifically investigates key elements that have been hypothesized to benefit from RL training.
  arXiv  Detail & Related papers  (2025-06-05T07:53:59Z)
• Alignment as Distribution Learning: Your Preference Model is Explicitly a Language Model [12.063078727764045]
  We argue that alignment via reinforcement learning from human feedback lacks theoretical justification and incentivizes deterministic solutions.<n>We propose three principled learning objectives: preference maximum likelihood estimation, preference distillation, and reverse KL minimization.<n>We empirically demonstrate that our distribution learning framework, especially preference distillation, consistently outperforms or matches the performances of RLHF and DPO.
  arXiv  Detail & Related papers  (2025-06-02T10:36:31Z)
• ProRL: Prolonged Reinforcement Learning Expands Reasoning Boundaries in Large Language Models [89.37819814048288]
  We introduce ProRL, a novel training methodology that incorporates KL divergence control, reference policy, and a diverse suite of tasks.<n>Our empirical analysis reveals that RL-trained models consistently outperform base resetting models across a wide range of pass@k evaluations.<n>These findings offer new insights into the conditions under which RL meaningfully expands reasoning boundaries in language models.
  arXiv  Detail & Related papers  (2025-05-30T17:59:01Z)
• The Limits of Preference Data for Post-Training [27.229909368242517]
  We show that preference data fundamentally and significantly limits outcome-based optimization.<n>We formalize this impossibility using voting theory, drawing an analogy between how a model chooses to answer a query with how voters choose a candidate to elect.<n>This indicates that grounded human scoring and algorithmic innovations are necessary for extending the success of RL post-training to domains demanding human feedback.
  arXiv  Detail & Related papers  (2025-05-26T13:26:15Z)
• Think or Not? Selective Reasoning via Reinforcement Learning for Vision-Language Models [67.87579664988199]
  TON is a two-stage training strategy for vision-language models (VLMs)<n>It introduces a think-or-not format that serves as a cold start for selective reasoning.<n>TON can reduce the completion length by up to 90% compared to vanilla GRPO.
  arXiv  Detail & Related papers  (2025-05-22T16:13:29Z)
• Is there Value in Reinforcement Learning? [1.534667887016089]
  Action-values play a central role in popular Reinforcement Learing (RL) models of behavior.<n>Critics had suggested that policy-gradient (PG) models should be favored over value-based (VB) ones.
  arXiv  Detail & Related papers  (2025-05-07T21:50:27Z)
• Do Theory of Mind Benchmarks Need Explicit Human-like Reasoning in Language Models? [14.29992535286614]
  Theory of Mind (ToM) is the ability to attribute mental states to others.<n>Recent advancements in Large Language Models have shown promising performance on ToM benchmarks.<n>Do these benchmarks necessitate explicit human-like reasoning processes, or can models succeed through alternative strategies?
  arXiv  Detail & Related papers  (2025-04-02T12:58:42Z)
• R-PRM: Reasoning-Driven Process Reward Modeling [53.06844294668382]
  Process Reward Models (PRMs) have emerged as a promising solution by evaluating each reasoning step.<n>Existing PRMs typically output evaluation scores directly, limiting both learning efficiency and evaluation accuracy.<n>We propose Reasoning-Driven Process Reward Modeling (R-PRM)<n>R-PRM generates seed data from limited annotations, effectively bootstrapping our model's reasoning capabilities.
  arXiv  Detail & Related papers  (2025-03-27T09:23:08Z)
• Stop Overthinking: A Survey on Efficient Reasoning for Large Language Models [54.04678363287392]
  Large Language Models (LLMs) have demonstrated remarkable capabilities in complex tasks.<n>Recent advancements in OpenAI o1 and DeepSeek-R1 have further improved performance in System-2 reasoning domains.
  arXiv  Detail & Related papers  (2025-03-20T17:59:38Z)
• Explore Theory of Mind: Program-guided adversarial data generation for theory of mind reasoning [88.68573198200698]
  We introduce ExploreToM, the first framework to allow large-scale generation of diverse and challenging theory of mind data.<n>Our approach leverages an A* search over a custom domain-specific language to produce complex story structures and novel, diverse, yet plausible scenarios.<n>Our evaluation reveals that state-of-the-art LLMs, such as Llama-3.1-70B and GPT-4o, show accuracies as low as 0% and 9% on ExploreToM-generated data.
  arXiv  Detail & Related papers  (2024-12-12T21:29:00Z)
• Pessimistic Model Selection for Offline Deep Reinforcement Learning [56.282483586473816]
  Deep Reinforcement Learning (DRL) has demonstrated great potentials in solving sequential decision making problems in many applications. One main barrier is the over-fitting issue that leads to poor generalizability of the policy learned by DRL. We propose a pessimistic model selection (PMS) approach for offline DRL with a theoretical guarantee.
  arXiv  Detail & Related papers  (2021-11-29T06:29:49Z)

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.