mlsanityData to 14 Oct 2026

Adaptive GoGI-Skip: Coupling Goal-Gradient Importance with Dynamic Uncertainty for Efficient Reasoning

arXiv 2505.08392 ↗First posted 13 May 2025cs.CL cs.AIcited by 12

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

Abstract (arXiv)

Chain-of-Thought (CoT) prompting trades inference speed for reasoning accuracy. Existing compressors force a compromise as static gradient techniques treat tokens independently, severing sequential logic, while uncertainty-based pruning ignores the final answer. We introduce Adaptive GoGI-Skip, a framework that resolves this tension by non-linearly coupling Goal-Gradient Importance (GoGI) with Adaptive Dynamic Skipping (ADS). GoGI quantifies each token's functional contribution to answer correctness via gradient sensitivity. ADS leverages runtime entropy to dynamically modulate the GoGI threshold, preserving low-gradient tokens essential for structural coherence at high-uncertainty junctions. Trained on 7,472 MATH traces, our policy transfers zero-shot to AIME, GPQA, and GSM8K, reducing token volume by >45% and accelerating inference up to 2.0× without accuracy loss. These results suggest that thinking-optimal compression demands synergy between teleological goals and epistemic uncertainty.

Key facts. Cited by 12 papers in the map (data through 14 Oct 2026). Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning. Builds on 36 of its 52 references that are in the map. First posted on arXiv 13 May 2025.

Builds on

36 of 52 references.

Closest ancestors · 13

Same direction, newest first

  1. 2025Reasoning Models Can Be Effective Without Thinking14 Apr
  2. 2025Concise Reasoning via Reinforcement Learning7 Apr
  3. 2025A Survey of Efficient Reasoning for Large Reasoning Models: Language, Multimodality, and Beyond27 Mar
  4. 2025DAST: Difficulty-Adaptive Slow-Thinking for Large Reasoning Models6 Mar
  5. 2025Towards Thinking-Optimal Scaling of Test-Time Compute for LLM Reasoning25 Feb
  6. 2025Reasoning with Latent Thoughts: On the Power of Looped Transformers24 Feb
  7. 2025TokenSkip: Controllable Chain-of-Thought Compression in LLMs17 Feb
  8. 2025Demystifying Long Chain-of-Thought Reasoning in LLMs5 Feb
  9. 2025O1-Pruner: Length-Harmonizing Fine-Tuning for O1-Like Reasoning Pruning22 Jan
  10. 2024C3oT: Generating Shorter Chain-of-Thought without Compromising Effectiveness16 Dec
  11. 2024Dualformer: Controllable Fast and Slow Thinking by Learning with Randomized Reasoning Traces13 Oct
  12. 2024Adaptive Inference-Time Compute: LLMs Can Predict if They Can Do Better, Even Mid-Generation3 Oct
1 more
  1. 2024Make Every Penny Count: Difficulty-Adaptive Self-Consistency for Cost-Efficient Reasoning24 Aug

Other references · 23

  1. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,380
  2. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,459
  3. 2021Training Verifiers to Solve Math Word Problemscited by 10,109
  4. 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,371
  5. 2024Qwen2.5 Technical Reportcited by 4,942
  6. 2017Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layercited by 3,509
  7. 2017Axiomatic Attribution for Deep Networkscited by 3,384
  8. 2023GPQA: A Graduate-Level Google-Proof Q&A Benchmarkcited by 2,879
  9. 2013Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Mapscited by 2,371
  10. 2020Transformers are RNNs: Fast Autoregressive Transformers with Linear Attentioncited by 2,022
  11. 2025Gemma 3 Technical Reportcited by 1,824
  12. 2025Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model?cited by 982
11 more
  1. 2019Attention is not Explanationcited by 929
  2. 2024Training Large Language Models to Reason in a Continuous Latent Spacecited by 797
  3. 2019Green AIcited by 627
  4. 2016Adaptive Computation Time for Recurrent Neural Networkscited by 506
  5. 2023Model Tells You What to Discard: Adaptive KV Cache Compression for LLMscited by 474
  6. 2024LLMLingua-2: Data Distillation for Efficient and Faithful Task-Agnostic Prompt Compressioncited by 348
  7. 2024On Designing Effective RL Reward at Training Time for LLM Reasoningcited by 87
  8. 2024LazyLLM: Dynamic Token Pruning for Efficient Long Context LLM Inferencecited by 70
  9. 2025Stepwise Perplexity-Guided Refinement for Efficient Chain-of-Thought Reasoning in Large Language Modelscited by 51
  10. 2024Gradient based Feature Attribution in Explainable AI: A Technical Reviewcited by 34
  11. 2025LLM-Microscope: Uncovering the Hidden Role of Punctuation in Context Memory of Transformerscited by 12

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 8; 2026: 4.

12 papers.

  1. 2025A Survey of LLM-Driven AI Agent Communication: Protocols, Security Risks, and Defense Countermeasurescited by 57
  2. 2025Don't Overthink It: A Survey of Efficient R1-style Large Reasoning Modelscited by 37
  3. 2025R1-Compress: Long Chain-of-Thought Compression via Chunk Compression and Searchcited by 30
  4. 2025Towards Concise and Adaptive Thinking in Large Reasoning Models: A Surveycited by 18
  5. 2025DRPO: Efficient Reasoning via Decoupled Reward Policy Optimizationcited by 16
  6. 2025SABER: Switchable and Balanced Training for Efficient LLM Reasoningcited by 12
  7. 2026Neural Chain-of-Thought Search: Searching the Optimal Reasoning Path to Enhance Large Language Modelscited by 8
  8. 2026VTC-R1: Vision-Text Compression for Efficient Long-Context Reasoningcited by 8
  9. 2025Reasoning Efficiently Through Adaptive Chain-of-Thought Compression: A Self-Optimizing Frameworkcited by 6
  10. 2025Focused Chain-of-Thought: Efficient LLM Reasoning via Structured Input Informationcited by 1
  11. 2026Bridging Efficiency and Transparency: Explainable CoT Compression in Multimodal Large Reasoning Modelscited by 0
  12. 2026Every Token Leaves a Ripple in the Stream of Thought: Eliciting Model-Internal Token Saliency for Chain-of-Thought Compressioncited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026ETR: Entropy Trend Reward for Efficient Chain-of-Thought Reasoning
  2. 2025Certainty-Guided Reasoning in Large Language Models: A Dynamic Thinking Budget Approach
  3. 2026EntroCut: Entropy-Guided Adaptive Truncation for Efficient Chain-of-Thought Reasoning in Small-scale Large Reasoning Models
  4. 2025MUR: Momentum Uncertainty guided Reasoning for Large Language Models
  5. 2025Entropy-Guided Loop: Achieving Reasoning through Uncertainty-Aware Generation
  6. 2026Improving reasoning at inference time via uncertainty minimisation
  7. 2026Free Energy Heuristics: Fast-And-Frugal Cognition as Active Inference Under Uncertain Precision
  8. 2026Memory Augmentation Unlocks Efficient Chain-of-Thought Reasoning