Imagen de Portada

Not All Correct Answers Are Equal: Why Your Distillation Source Matters

Fuente: arXiv

Guardado en:

Detalles Bibliográficos
Autores principales:	Tian, Xiaoyu, Ji, Yunjie, Wang, Haotian, Chen, Shuaiting, Zhao, Sitong, Peng, Yiping, Zhao, Han, Li, Xiangang
Formato:	Preprint
Publicado:	2025
Materias:	Computation and Language
Acceso en línea:	Acceder al recurso
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

Ejemplares similares

1.4 Million Open-Source Distilled Reasoning Dataset to Empower Large Language Model Training
por: Zhao, Han, et al.
Publicado: (2025)

Leveraging Reasoning Model Answers to Enhance Non-Reasoning Model Capability
por: Wang, Haotian, et al.
Publicado: (2025)

DeepDistill: Enhancing LLM Reasoning Capabilities via Large-Scale Difficulty-Graded Data Training
por: Tian, Xiaoyu, et al.
Publicado: (2025)

AM-Thinking-v1: Advancing the Frontier of Reasoning at 32B Scale
por: Ji, Yunjie, et al.
Publicado: (2025)

Think Twice: Enhancing LLM Reasoning by Scaling Multi-round Test-time Thinking
por: Tian, Xiaoyu, et al.
Publicado: (2025)

How Difficulty-Aware Staged Reinforcement Learning Enhances LLMs' Reasoning Capabilities: A Preliminary Experimental Study
por: Ji, Yunjie, et al.
Publicado: (2025)

Exploring the Potential of Offline RL for Reasoning in LLMs: A Preliminary Study
por: Tian, Xiaoyu, et al.
Publicado: (2025)

Not All Proofs Are Equal: Evaluating LLM Proof Quality Beyond Correctness
por: Petrov, Ivo, et al.
Publicado: (2026)

Not All Parameters Are Created Equal: Smart Isolation Boosts Fine-Tuning Performance
por: Wang, Yao, et al.
Publicado: (2025)

Not All Contexts Are Equal: Teaching LLMs Credibility-aware Generation
por: Pan, Ruotong, et al.
Publicado: (2024)

Beyond Answers: Transferring Reasoning Capabilities to Smaller LLMs Using Multi-Teacher Knowledge Distillation
por: Tian, Yijun, et al.
Publicado: (2024)

Mitigating Spurious Correlations Between Question and Answer via Chain-of-Thought Correctness Perception Distillation
por: Xie, Hongyan, et al.
Publicado: (2025)

Not All Languages are Equal: Insights into Multilingual Retrieval-Augmented Generation
por: Wu, Suhang, et al.
Publicado: (2024)

Memorizing is Not Enough: Deep Knowledge Injection Through Reasoning
por: Xu, Ruoxi, et al.
Publicado: (2025)

Answer is All You Need: Instruction-following Text Embedding via Answering the Question
por: Peng, Letian, et al.
Publicado: (2024)

Not All Errors Are Created Equal: ASCoT Addresses Late-Stage Fragility in Efficient LLM Reasoning
por: Zhang, Dongxu, et al.
Publicado: (2025)

All Claims Are Equal, but Some Claims Are More Equal Than Others: Importance-Sensitive Factuality Evaluation of LLM Generations
por: Wanner, Miriam, et al.
Publicado: (2025)

Sandwich Reasoning: An Answer-Reasoning-Answer Approach for Low-Latency Query Correction
por: Zhang, Chen, et al.
Publicado: (2026)

Evaluating and Calibrating LLM Confidence on Questions with Multiple Correct Answers
por: Wang, Yuhan, et al.
Publicado: (2026)

Not All Synthetic Data Is Yours to Learn From
por: Alemohammad, Sina, et al.
Publicado: (2026)

LexPro-1.0 Technical Report
por: Chen, Haotian, et al.
Publicado: (2025)

ASTRA: Automated Synthesis of agentic Trajectories and Reinforcement Arenas
por: Tian, Xiaoyu, et al.
Publicado: (2026)

MMBench: Is Your Multi-modal Model an All-around Player?
por: Liu, Yuan, et al.
Publicado: (2023)

SARI: Structured Audio Reasoning via Curriculum-Guided Reinforcement Learning
por: Wen, Cheng, et al.
Publicado: (2025)

Your Co-Workers Matter: Evaluating Collaborative Capabilities of Language Models in Blocks World
por: Wu, Guande, et al.
Publicado: (2024)

Every Answer Matters: Evaluating Commonsense with Probabilistic Measures
por: Cheng, Qi, et al.
Publicado: (2024)

Seeing Isn't Knowing: Do VLMs Know When Not to Answer Spatial Questions (and Why)?
por: Zhang, Yue, et al.
Publicado: (2026)

LLM-based Privacy Data Augmentation Guided by Knowledge Distillation with a Distribution Tutor for Medical Text Classification
por: Song, Yiping, et al.
Publicado: (2024)

Enhancing Multimodal Sentiment Analysis for Missing Modality through Self-Distillation and Unified Modality Cross-Attention
por: Weng, Yuzhe, et al.
Publicado: (2024)

LLMs Can Generate a Better Answer by Aggregating Their Own Responses
por: Li, Zichong, et al.
Publicado: (2025)

All Entities are Not Created Equal: Examining the Long Tail for Ultra-Fine Entity Typing
por: Deshmukh, Advait, et al.
Publicado: (2024)

Not All Errors Are Equal: Investigation of Speech Recognition Errors in Alzheimer's Disease Detection
por: Kang, Jiawen, et al.
Publicado: (2024)

Is That Your Final Answer? Test-Time Scaling Improves Selective Question Answering
por: Jurayj, William, et al.
Publicado: (2025)

Not All Thoughts are Generated Equal: Efficient LLM Reasoning via Multi-Turn Reinforcement Learning
por: Ning, Yansong, et al.
Publicado: (2025)

Not All LLM-Generated Data Are Equal: Rethinking Data Weighting in Text Classification
por: Kuo, Hsun-Yu, et al.
Publicado: (2024)

Same Question, Different Source, Different Answer: Auditing Source-Dependence in Medical Multi-Source RAG
por: Li, Yubo, et al.
Publicado: (2026)

Not All Uncertainty Is Equal: How Uncertainty Granularity Shapes Human Verification in LLM-Assisted Decision Making
por: Villavicencio, Mauricio, et al.
Publicado: (2026)

Correct Answers from Sound Reasoning: Verifiable Process Supervision for Language Models
por: Kim, Kyuyoung, et al.
Publicado: (2026)

Your Teacher Can't Help You Here: Combating Supervision Fidelity Decay in On-Policy Distillation
por: Liu, Yanjiang, et al.
Publicado: (2026)

Not All Preference Pairs Are Created Equal: A Recipe for Annotation-Efficient Iterative Preference Learning
por: Yang, Sen, et al.
Publicado: (2024)