Just Enough Thinking: Efficient Reasoning with Adaptive Length Penalties Reinforcement Learning

• URL: http://arxiv.org/abs/2506.05256v2
• Date: Fri, 06 Jun 2025 02:38:39 GMT
• Title: Just Enough Thinking: Efficient Reasoning with Adaptive Length Penalties Reinforcement Learning
• Authors: Violet Xiang, Chase Blagden, Rafael Rafailov, Nathan Lile, Sang Truong, Chelsea Finn, Nick Haber,
• Abstract summary: Posttraining DeepScaleR-1.5B with ALP cuts average token usage by 50% without significantly dropping performance.<n>Relative to fixed-budget and uniform penalty baselines, ALP redistributes its reduced budget more intelligently by cutting compute on easy prompts and reallocating saved tokens to difficult ones.
• Score: 42.82825782517565
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large reasoning models (LRMs) achieve higher performance on challenging reasoning tasks by generating more tokens at inference time, but this verbosity often wastes computation on easy problems. Existing solutions, including supervised finetuning on shorter traces, user-controlled budgets, or RL with uniform penalties, either require data curation, manual configuration, or treat all problems alike regardless of difficulty. We introduce Adaptive Length Penalty (ALP), a reinforcement learning objective tailoring generation length to per-prompt solve rate. During training, ALP monitors each prompt's online solve rate through multiple rollouts and adds a differentiable penalty whose magnitude scales inversely with that rate, so confident (easy) prompts incur a high cost for extra tokens while hard prompts remain unhindered. Posttraining DeepScaleR-1.5B with ALP cuts average token usage by 50\% without significantly dropping performance. Relative to fixed-budget and uniform penalty baselines, ALP redistributes its reduced budget more intelligently by cutting compute on easy prompts and reallocating saved tokens to difficult ones, delivering higher accuracy on the hardest problems with higher cost.

Related papers

• e1: Learning Adaptive Control of Reasoning Effort [88.51897900019485]
  Increasing the thinking budget of AI models can significantly improve accuracy, but not all questions warrant the same amount of reasoning.<n>Users may prefer to allocate different amounts of reasoning effort depending on how they value output quality versus latency and cost.<n>We propose Adaptive Effort Control, a self-adaptive reinforcement learning method that trains models to use a user-specified fraction of tokens.
  arXiv  Detail & Related papers  (2025-10-30T23:12:21Z)
• DLER: Doing Length pEnalty Right - Incentivizing More Intelligence per Token via Reinforcement Learning [134.03095505580276]
  Doing Length pEnalty Right (DLER) is a training recipe combining batch-wise reward normalization, higher clipping, dynamic sampling, and a simple truncation length penalty.<n>DLER achieves state-of-the-art accuracy--efficiency trade-offs, cutting output length by over 70 percent while surpassing all previous baseline accuracy.
  arXiv  Detail & Related papers  (2025-10-16T20:05:57Z)
• Intra-request branch orchestration for efficient LLM reasoning [52.68946975865865]
  Large Language Models (LLMs) increasingly rely on inference-time reasoning algorithms to improve accuracy on complex tasks.<n>Prior work has largely focused on reducing token usage, often at the expense of accuracy, while overlooking other latency factors.<n>We present DUCHESS, an LLM serving system that reduces cost and latency without sacrificing accuracy through intra-request branch orchestration guided by predictions.
  arXiv  Detail & Related papers  (2025-09-29T15:52:08Z)
• Sample More to Think Less: Group Filtered Policy Optimization for Concise Reasoning [7.260825775935882]
  Group Filtered Policy Optimization curbs this length explosion by sampling larger groups per problem during training.<n>GFPO cuts GRPO's length inflation by 46-71% across challenging STEM and coding benchmarks.<n> Adaptive Difficulty GFPO allocates more training resources to harder problems based on real-time difficulty estimates.
  arXiv  Detail & Related papers  (2025-08-13T11:43:49Z)
• R-Stitch: Dynamic Trajectory Stitching for Efficient Reasoning [80.104336426172]
  Chain-of-thought (CoT) enhances problem-solving ability of large language models.<n>CoT incurs substantial inference cost due to long autoregressive trajectories.<n>We introduce R-Stitch, a training-free hybrid decoding framework.
  arXiv  Detail & Related papers  (2025-07-23T08:14:36Z)
• Steering LLM Thinking with Budget Guidance [48.65894557568655]
  Budget guidance is a method for steering the reasoning process of LLMs toward a target budget without requiring any fine-tuning.<n>Our approach introduces a lightweight predictor that models a Gamma distribution over the remaining thinking length.<n>This signal is then used to guide generation in a soft, token-level manner, ensuring that the overall reasoning trace adheres to the specified thinking budget.
  arXiv  Detail & Related papers  (2025-06-16T17:57:05Z)
• Improving Data Efficiency for LLM Reinforcement Fine-tuning Through Difficulty-targeted Online Data Selection and Rollout Replay [61.823835392216544]
  Reinforcement learning (RL) has become an effective approach for fine-tuning large language models (LLMs)<n>We propose two techniques to improve data efficiency in LLM RL fine-tuning: difficulty-targeted online data selection and rollout replay.<n>Our method reduces RL fine-tuning time by 25% to 65% to reach the same level of performance as the original GRPO algorithm.
  arXiv  Detail & Related papers  (2025-06-05T17:55:43Z)
• TACO: Think-Answer Consistency for Optimized Long-Chain Reasoning and Efficient Data Learning via Reinforcement Learning in LVLMs [50.820065021136024]
  DeepSeek R1 has significantly advanced complex reasoning for large language models (LLMs)<n>Recent methods have attempted to replicate R1's reasoning capabilities in multimodal settings.<n>We propose TACO, a novel reinforcement learning algorithm for visual reasoning.
  arXiv  Detail & Related papers  (2025-05-27T06:30:48Z)
• AdaCtrl: Towards Adaptive and Controllable Reasoning via Difficulty-Aware Budgeting [23.004467211806467]
  AdaCtrl is a novel framework to support difficulty-aware adaptive reasoning budget allocation.<n>It dynamically adjusts its reasoning length based on self-assessed problem difficulty.<n>AdaCtrl enables precise user control over the reasoning budget, allowing for tailored responses to meet specific needs.
  arXiv  Detail & Related papers  (2025-05-24T18:46:50Z)
• Thinking Fast and Right: Balancing Accuracy and Reasoning Length with Adaptive Rewards [17.829990749622496]
  We propose an adaptive reward-shaping method for large language models.<n>Our method dynamically adjusts the trade-off between accuracy and response length based on model performance.<n> Experiments show that our approach consistently and dramatically reduces reasoning length while largely maintaining accuracy.
  arXiv  Detail & Related papers  (2025-05-23T18:44:46Z)
• SelfBudgeter: Adaptive Token Allocation for Efficient LLM Reasoning [29.64638547097158]
  SelfBudgeter is a self-adaptive controllable reasoning strategy for efficient reasoning.<n>We introduce budget-guided GPRO for reinforcement learning, which effectively maintains accuracy while reducing output length.<n> Experimental results demonstrate that SelfBudgeter can rationally allocate budgets according to problem complexity.
  arXiv  Detail & Related papers  (2025-05-16T14:08:04Z)
• Sparsity Forcing: Reinforcing Token Sparsity of MLLMs [40.93786579652003]
  We explicitly reinforce token sparsity in well-posed multimodal large language models (MLLMs) through a simple RL-based post-training framework named textitSparsity Forcing.<n>Our method explores the efficiency-accuracy trade-off by running multiple rollouts with different token budgets, where both efficiency (token reduction ratio) and performance (answer correctness) are formulated as joint rewards.
  arXiv  Detail & Related papers  (2025-04-23T01:45:55Z)
• DAST: Difficulty-Adaptive Slow-Thinking for Large Reasoning Models [30.184895117009457]
  This paper introduces Difficulty-Adaptive Slow Thinking (DAST), a novel framework that enables models to autonomously adjust the length of Chain-of-Thought (CoT) based on problem difficulty.<n>Experiments on diverse datasets and model scales demonstrate that DAST effectively mitigates overthinking while preserving reasoning accuracy on complex problems.
  arXiv  Detail & Related papers  (2025-03-06T14:23:06Z)
• Self-Regulation and Requesting Interventions [63.5863047447313]
  We propose an offline framework that trains a "helper" policy to request interventions.<n>We score optimal intervention timing with PRMs and train the helper model on these labeled trajectories.<n>This offline approach significantly reduces costly intervention calls during training.
  arXiv  Detail & Related papers  (2025-02-07T00:06:17Z)
• Switching the Loss Reduces the Cost in Batch (Offline) Reinforcement Learning [57.154674117714265]
  We show that the number of samples needed to learn a near-optimal policy with FQI-log scales with the accumulated cost of the optimal policy. We empirically verify that FQI-log uses fewer samples than FQI trained with squared loss on problems where the optimal policy reliably achieves the goal.
  arXiv  Detail & Related papers  (2024-03-08T15:30:58Z)

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.