Image de couverture de livre

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Zhang, Yunxiang, Khalifa, Muhammad, Zhang, Lechen, Liu, Xin, Lee, Ayoung, Zhang, Xinliang Frederick, Bayat, Farima Fatahi, Wang, Lu
Format:	Preprint
Publié:	2025
Sujets:	Computation and Language
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

Logit Arithmetic Elicits Long Reasoning Capabilities Without Training
par: Zhang, Yunxiang, et autres
Publié: (2025)

FactBench: A Dynamic Benchmark for In-the-Wild Language Model Factuality Evaluation
par: Bayat, Farima Fatahi, et autres
Publié: (2024)

Enhancing Language Model Factuality via Activation-Based Confidence Calibration and Guided Decoding
par: Liu, Xin, et autres
Publié: (2024)

From Proof to Program: Characterizing Tool-Induced Reasoning Hallucinations in Large Language Models
par: Bayat, Farima Fatahi, et autres
Publié: (2025)

Enhanced Language Model Truthfulness with Learnable Intervention and Uncertainty Expression
par: Bayat, Farima Fatahi, et autres
Publié: (2024)

Skill-Aware Data Selection and Fine-Tuning for Data-Efficient Reasoning Distillation
par: Zhang, Lechen, et autres
Publié: (2026)

TSUBASA: Improving Long-Horizon Personalization via Evolving Memory and Self-Learning with Context Distillation
par: Zhang, Xinliang Frederick, et autres
Publié: (2026)

Small Language Models Need Strong Verifiers to Self-Correct Reasoning
par: Zhang, Yunxiang, et autres
Publié: (2024)

LitCab: Lightweight Language Model Calibration over Short- and Long-form Responses
par: Liu, Xin, et autres
Publié: (2023)

Narrative-of-Thought: Improving Temporal Reasoning of Large Language Models via Recounted Narratives
par: Zhang, Xinliang Frederick, et autres
Publié: (2024)

VeriFact: Enhancing Long-Form Factuality Evaluation with Refined Fact Extraction and Reference Facts
par: Liu, Xin, et autres
Publié: (2025)

LiveOIBench: Can Large Language Models Outperform Human Contestants in Informatics Olympiads?
par: Zou, Kaijian, et autres
Publié: (2025)

PRIME: Large Language Model Personalization with Cognitive Dual-Memory and Personalized Thought Process
par: Zhang, Xinliang Frederick, et autres
Publié: (2025)

On Many-Shot In-Context Learning for Long-Context Evaluation
par: Zou, Kaijian, et autres
Publié: (2024)

GRACE: Discriminator-Guided Chain-of-Thought Reasoning
par: Khalifa, Muhammad, et autres
Publié: (2023)

Causal Language Modeling Can Elicit Search and Reasoning Capabilities on Logic Puzzles
par: Shah, Kulin, et autres
Publié: (2024)

MOKA: Moral Knowledge Augmentation for Moral Event Extraction
par: Zhang, Xinliang Frederick, et autres
Publié: (2023)

Gaming the Judge: Unfaithful Chain-of-Thought Can Undermine Agent Evaluation
par: Khalifa, Muhammad, et autres
Publié: (2026)

APE: Active Learning-based Tooling for Finding Informative Few-shot Examples for LLM-based Entity Matching
par: Qian, Kun, et autres
Publié: (2024)

An Empirical Study on Eliciting and Improving R1-like Reasoning Models
par: Chen, Zhipeng, et autres
Publié: (2025)

Unlocking General Long Chain-of-Thought Reasoning Capabilities of Large Language Models via Representation Engineering
par: Tang, Xinyu, et autres
Publié: (2025)

API Is Enough: Conformal Prediction for Large Language Models Without Logit-Access
par: Su, Jiayuan, et autres
Publié: (2024)

How Numerical Precision Affects Arithmetical Reasoning Capabilities of LLMs
par: Feng, Guhao, et autres
Publié: (2024)

Reasoning Vectors: Transferring Chain-of-Thought Capabilities via Task Arithmetic
par: Zbeeb, Mohammad, et autres
Publié: (2025)

When Disagreements Elicit Robustness: Investigating Self-Repair Capabilities under LLM Multi-Agent Disagreements
par: Ju, Tianjie, et autres
Publié: (2025)

Eliciting In-context Retrieval and Reasoning for Long-context Large Language Models
par: Qiu, Yifu, et autres
Publié: (2025)

VersaTune: An Efficient Data Composition Framework for Training Multi-Capability LLMs
par: Lu, Keer, et autres
Publié: (2024)

Elicit and Enhance: Advancing Multimodal Reasoning in Medical Scenarios
par: Huang, Zhongzhen, et autres
Publié: (2025)

Source-Aware Training Enables Knowledge Attribution in Language Models
par: Khalifa, Muhammad, et autres
Publié: (2024)

AlchemistCoder: Harmonizing and Eliciting Code Capability by Hindsight Tuning on Multi-source Data
par: Song, Zifan, et autres
Publié: (2024)

LogitsCoder: Towards Efficient Chain-of-Thought Path Search via Logits Preference Decoding for Code Generation
par: Chen, Jizheng, et autres
Publié: (2026)

Are Your LLMs Capable of Stable Reasoning?
par: Liu, Junnan, et autres
Publié: (2024)

Think Through Uncertainty: Improving Long-Form Generation Factuality via Reasoning Calibration
par: Liu, Xin, et autres
Publié: (2026)

Evaluating Accounting Reasoning Capabilities of Large Language Models
par: Zhou, Jie, et autres
Publié: (2026)

Logic-Regularized Verifier Elicits Reasoning from LLMs
par: Wang, Xinyu, et autres
Publié: (2026)

Paraphrase Types Elicit Prompt Engineering Capabilities
par: Wahle, Jan Philip, et autres
Publié: (2024)

RAT: Retrieval Augmented Thoughts Elicit Context-Aware Reasoning in Long-Horizon Generation
par: Wang, Zihao, et autres
Publié: (2024)

Case Study: Testing Model Capabilities in Some Reasoning Tasks
par: Zhang, Min, et autres
Publié: (2024)

Divide, Reweight, and Conquer: A Logit Arithmetic Approach for In-Context Learning
par: Huang, Chengsong, et autres
Publié: (2024)

Diversity of Thought Elicits Stronger Reasoning Capabilities in Multi-Agent Debate Frameworks
par: Hegazy, Mahmood
Publié: (2024)