mlsanityData to 14 Oct 2026

SPREG: Structured Plan Repair with Entropy-Guided Test-Time Intervention for Large Language Model Reasoning

arXiv 2604.17884 ↗First posted 20 Apr 2026cs.AIcited by 0

On the map: LLM reasoning → Test-time compute and efficient thinking

Abstract (arXiv)

Large Language Models (LLMs) are prone to logical hallucinations and stochastic drifts during long-chain reasoning. While Classifier-Free Guidance (CFG) can improve instruction adherence, standard static implementations often cause semantic dilution and linguistic degradation. We propose SPREG (Structured Plan-guided Real-time Entropy Gating), a lightweight inference-time framework for surgical error rectification. SPREG employs an adaptive dual-threshold mechanism to monitor real-time entropy, identifying sudden ``entropy spikes'' as reliable indicators of logical failure. Upon detection, it triggers a dynamic repair by replacing uninformative null-priors with reference distributions synthesized from historical high-confidence states. By modulating guidance intensity according to structured reasoning stages (e.g., Action, Observation), SPREG steers the model back to a stable manifold without compromising fluency. Our experiments demonstrate significant gains, notably a 20.0% absolute accuracy improvement on AIME25, while effectively suppressing uncontrolled entropy drift in complex tasks.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning. Builds on 23 of its 25 references that are in the map. First posted on arXiv 20 Apr 2026.

Builds on

23 of 25 references.

Closest ancestors · 2

Same direction, newest first

  1. 2023Tree of Thoughts: Deliberate Problem Solving with Large Language Models17 May
  2. 2022Language Models (Mostly) Know What They Know11 Jul

Other references · 21

  1. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,459
  2. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,297
  3. 2025Qwen3 Technical Reportcited by 8,820
  4. 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,695
  5. 2022Classifier-Free Diffusion Guidancecited by 5,847
  6. 2022Self-Consistency Improves Chain of Thought Reasoning in Language Modelscited by 5,708
  7. 2015Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learningcited by 4,448
  8. 2023Let's Verify Step by Stepcited by 3,789
  9. 2023GPQA: A Graduate-Level Google-Proof Q&A Benchmarkcited by 2,879
  10. 2022Survey of Hallucination in Natural Language Generationcited by 2,075
  11. 2022Least-to-Most Prompting Enables Complex Reasoning in Large Language Modelscited by 1,407
  12. 2024OlympiadBench: A Challenging Benchmark for Promoting AGI with Olympiad-Level Bilingual Multimodal Scientific Problemscited by 1,136
9 more
  1. 2021Show Your Work: Scratchpads for Intermediate Computation with Language Modelscited by 886
  2. 2023Semantic Uncertainty: Linguistic Invariances for Uncertainty Estimation in Natural Language Generationcited by 819
  3. 2022Contrastive Decoding: Open-ended Text Generation as Optimizationcited by 609
  4. 2023DoLa: Decoding by Contrasting Layers Improves Factuality in Large Language Modelscited by 376
  5. 2024ReST-MCTS*: LLM Self-Training via Process Reward Guided Tree Searchcited by 344
  6. 2024DPO Meets PPO: Reinforced Token Optimization for RLHFcited by 127
  7. 2023Stay on topic with Classifier-Free Guidancecited by 91
  8. 2025Agentic Reward Modeling: Integrating Human Preferences with Verifiable Correctness Signals for Reliable Reward Systemscited by 51
  9. 2025DaGRPO: Rectifying Gradient Conflict in Reasoning via Distinctiveness-Aware Group Relative Policy Optimizationcited by 9

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026REWIND: Uncertainty-Guided Trajectory Repair for Test-Time Reasoning
  2. 2025Think Just Enough: Sequence-Level Entropy as a Confidence Signal for LLM Reasoning
  3. 2025Adaptive Termination for Multi-round Parallel Reasoning: An Universal Semantic Entropy-Guided Framework
  4. 2026Dissecting Failure Dynamics in Large Language Model Reasoning
  5. 2026The Shape of Overthinking: Backtracking Bursts in Long Reasoning Traces
  6. 2026Latent Phase-Shift Rollback: Inference-Time Error Correction via Residual Stream Monitoring and KV-Cache Steering
  7. 2026ENTRA: Entropy-Based Redundancy Avoidance in Large Language Model Reasoning
  8. 2026ReTreVal: Reasoning Tree with Validation and Cross-Problem Memory for Large Language Models