CritICL省推理

Papers arxiv:2608.

27455 Copy markdown CritICL: Inference-Time Weak-to-Strong Generalization from Small Language Model Failure Modes Published on Aug 27 · Submitted by Yinghui He on Aug 28 Upvote 9 +1 Authors: Yufan Wu ,Yinghui He ,Zhengyi Hu ,Lang Wei ,Ruichen Li ,Qifan Yang ,Ting Zhu Abstract CritICL improves LLM reasoning at inference time by using structured failure patterns from weaker models as critique-based guidance, reducing generation and token costs.

Generated by thinkingmachines/Inkling-Small Recent advances in inference-time scaling have significantly improved the reasoning performance of large language models (LLMs).However, these methods typically rely on repeated generation or external verification.

To address this limitation, we introduce CritICL, a novel inference-time framework that improves reasoning while maintaining high efficiency.Our key insight is that LLM failure modes exhibit structured patterns across model scales within the same family.

Instead of treating failures as undesirable outputs, CritICL leverages them as a source of guidance.Specifically, we utilize failure modes derived from weaker models and incorporate them into inference through critique-based in-context examples.

We propose two variants: CritICL-dynamic, which adaptively predicts input-specific failure modes and retrieves critiques, and CritICL-static, which uses a global failure mode profile to provide stable guidance.

Experimental results show that CritICL consistently outperforms standard in-context learning and achieves performance competitive with or superior to test-time scaling methods, while requiring significantly fewer generations and lower token cost.Code available at: https://github.

com/umwyf/CRITICL View arXiv page View PDF Add to collection Community yinghuihe Paper submitter 2 days ago Recent advances in inference-time scaling have significantly improved the reasoning performance of large language models (LLMs).

However, these methods typically rely on repeated generation or external verification.To address this limitation, we introduce CritICL, a novel inference-time framework that improves reasoning while maintaining high efficiency.

Our key insight is that LLM failure modes exhibit structured patterns across model scales within the same family.Instead of treating failures as undesirable outputs, CritICL leverages them as a source of guidance.

Specifically, we utilize failure modes derived from weaker models and incorporate them into inference through critique-based in-context examples.

We propose two variants: CritICL-dynamic, which adaptively predicts input-specific failure modes and retrieves critiques, and CritICL-static, which uses a global failure mode profile to provide stable guidance.

Experimental results show that CritICL consistently outperforms standard in-context learning and achieves performance competitive with or superior to test-time scaling methods, while requiring significantly fewer generations and lower token cost.

Reply librarian-bot 2 days ago This is an automated message from the Librarian Bot.I found the following papers similar to this paper.

The following papers were recommended by the Semantic Scholar API Learning from Synthetic Data without Model Collapse in Iterative Instruction Tuning (2026) ThinkRetrieve: Retrieval-Augmented Reasoning Traces for Test-Time Scaling (2026) Refining Over Resampling: Test-Time Self-Correction for LLM Reasoning (2026) ReflectRL: Learning from Golden Negative Trajectories via Reflective-to-Direct Reasoning (2026) StrategyBench: Evaluating Explicit Strategy Induction in Large Language Models (2026) AI4AI at Test-Time: Strong-to-Weak Capability Transfer via Harnesses (2026) SPARK: Susceptibility-Guided Profiling and Steering of Latent Reasoning States in Large Language Models (2026) Please give a thumbs up to this comment if you found it helpful!

If you want recommendations for any Paper on Hugging Face checkout this Space You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: @librarian-bot recommend Reply EditPreview Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images Comment · Sign up or log in to comment Upvote 9 Get this paper in your agent: hf papers read 2608.27455 Don't have the latest CLI?curl -LsSf https://hf.co/cli/install.sh | bash Models citing this paper 0 No model linking this paper Cite arxiv.org/abs/2608.

27455 in a model README.md to link it from this page.Datasets citing this paper 0 No dataset linking this paper Cite arxiv.org/abs/2608.27455 in a dataset README.md to link it from this page.Spaces citing this paper 0 No Space linking this paper Cite arxiv.org/abs/2608.27455 in a Space README.

md to link it from this page.Collections including this paper 1