LLMs Gaming Verifiers: RLVR can Lead to Reward Hacking

• URL: http://arxiv.org/abs/2604.15149v1
• Date: Thu, 16 Apr 2026 15:30:10 GMT
• Title: LLMs Gaming Verifiers: RLVR can Lead to Reward Hacking
• Abstract summary: We study inductive reasoning tasks where models must induce and output logical rules.<n>We find that RLVR-trained models systematically abandon rule induction.<n>We show that this behavior is not a failure of understanding but a form of reward hacking.
• Score: 43.633688879582714
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: As reinforcement Learning with Verifiable Rewards (RLVR) has become the dominant paradigm for scaling reasoning capabilities in LLMs, a new failure mode emerges: LLMs gaming verifiers. We study this phenomenon on inductive reasoning tasks, where models must induce and output logical rules. We find that RLVR-trained models systematically abandon rule induction. Instead of learning generalizable patterns (e.g., ``trains carrying red cars go east''), they enumerate instance-level labels, producing outputs that pass verifiers without capturing the relational patterns required by the task. We show that this behavior is not a failure of understanding but a form of reward hacking: imperfect verifiers that check only extensional correctness admit false positives. To detect such shortcuts, we introduce Isomorphic Perturbation Testing (IPT), which evaluates a single model output under both extensional and isomorphic verification, where the latter enforces invariance under logically isomorphic tasks. While genuine rule induction remains invariant, shortcut strategies fail. We find that shortcut behavior is specific to RLVR-trained reasoning models (e.g., GPT-5, Olmo3) and absent in non-RLVR models (e.g., GPT-4o, GPT-4.5, Ministral). Moreover, shortcut prevalence increases with task complexity and inference-time compute. In controlled training experiments, extensional verification directly induces shortcut strategies, while isomorphic verification eliminates them. These results show that RLVR can incentivize reward hacking not only through overt manipulation but also by exploiting what the verifier fails to enforce.

Related papers

• From RLVR to RLSVR: Task Transformation Induces Self-Verifiable Rewards for Open-Ended LLM Self-Improvement [56.26277769790695]
  Reinforcement Learning with Verifiable Rewards (RLVR) has driven recent progress in reasoning-oriented large language models (LLMs)<n>We propose Reinforcement Learning with Self-Verifiable Rewards (RLSVR), a task-transformation-based training paradigm for extending RLVR to open-ended tasks.<n>We instantiate RLSVR with SpyRL, a Self-PlaY Reinforcement Learning method inspired by social deduction game Who Is the Spy?
  arXiv  Detail & Related papers  (2026-07-26T19:04:33Z)
• When Reward Hacking Rebounds: Understanding and Mitigating It with Representation-Level Signals [11.280037154530847]
  Reinforcement learning for LLMs is vulnerable to reward hacking.<n>We study this phenomenon in coding tasks using an environment-manipulation setting.<n>We propose Advantage Modification, which integrates shortcut concept scores into GRPO advantage computation.
  arXiv  Detail & Related papers  (2026-04-01T23:33:08Z)
• How Far Can Unsupervised RLVR Scale LLM Training? [57.44753418846446]
  Unsupervised reinforcement learning with verifiable rewards (URLVR) offers a pathway to scale LLM training beyond the supervision bottleneck.<n>Recent works leverage model intrinsic signals, showing promising early gains, yet their potential and limitations remain unclear.<n>We classify URLVR methods into intrinsic versus external based on reward sources, then establish a unified theoretical framework.
  arXiv  Detail & Related papers  (2026-03-09T17:38:11Z)
• From Verifiable Dot to Reward Chain: Harnessing Verifiable Reference-based Rewards for Reinforcement Learning of Open-ended Generation [52.62655622099456]
  We propose reinforcement learning with verifiable reference-based rewards (RLVRR)<n>Instead of checking the final answer, RLVRR extracts an ordered linguistic signal from high-quality references (i.e., reward chain)<n>In this way, RLVRR decomposes rewards into two dimensions: content, which preserves deterministic core concepts, and style, which evaluates adherence to stylistic properties.
  arXiv  Detail & Related papers  (2026-01-26T14:39:58Z)
• Spurious Rewards Paradox: Mechanistically Understanding How RLVR Activates Memorization Shortcuts in LLMs [16.74831908818562]
  Recent evidence shows models like Qwen 2.5 achieve significant gains even with spurious or incorrect rewards.<n>We identify a "Perplexity Paradox": spurious RLVR triggers a divergence where answer-token perplexity drops while prompt-side coherence degrades.<n>We uncover a hidden Anchor-Adapter circuit that facilitates this shortcut.
  arXiv  Detail & Related papers  (2026-01-16T07:55:38Z)
• Boosting Reasoning in Large Multimodal Models via Activation Replay [136.6522463570943]
  We show how RLVR shifts low-entropy activations unexpectedly, while high-entropy ones are less affected.<n>We propose Activation Replay, a training-free approach that boosts multimodal reasoning of post-trained LMMs.
  arXiv  Detail & Related papers  (2025-11-25T06:31:57Z)
• The Ends Justify the Thoughts: RL-Induced Motivated Reasoning in LLMs [2.583082967853897]
  We find that motivated reasoning can be detected by most frontier reasoning models.<n>We find that as models become more sophisticated, their motivated reasoning may become increasingly difficult for monitors to detect.
  arXiv  Detail & Related papers  (2025-10-20T00:24:08Z)
• The Surprising Effectiveness of Negative Reinforcement in LLM Reasoning [37.13807960501503]
  Reinforcement learning with verifiable rewards (RLVR) is a promising approach for training language models (LMs)<n>We decompose the learning signal into reinforcing correct responses and penalizing incorrect ones, referred to as Positive and Negative Sample Reinforcement (PSR and NSR)<n>We show that NSR works by suppressing incorrect generations and redistributing probability mass toward other plausible candidates, guided by the model's prior beliefs.
  arXiv  Detail & Related papers  (2025-06-02T06:10:54Z)
• GTR: Guided Thought Reinforcement Prevents Thought Collapse in RL-based VLM Agent Training [62.536191233049614]
  Reinforcement learning with verifiable outcome rewards (RLVR) has effectively scaled up chain-of-thought (CoT) reasoning in large language models (LLMs)<n>This work investigates this problem through extensive experiments on complex card games, such as 24 points, and embodied tasks from ALFWorld.<n>We find that when rewards are based solely on action outcomes, RL fails to incentivize CoT reasoning in VLMs, instead leading to a phenomenon we termed thought collapse.
  arXiv  Detail & Related papers  (2025-03-11T15:17:02Z)

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.