Cover Image

Quantization Meets Reasoning: Exploring and Mitigating Degradation of Low-Bit LLMs in Mathematical Reasoning

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Li, Zhen, Su, Yupeng, Wang, Songmiao, Yang, Runming, Xie, Congkai, Liu, Aofan, Li, Ming, Cao, Jiannong, Xie, Yuan, Wong, Ngai, Yang, Hongxia
Format:	Preprint
Published:	2025
Subjects:	Machine Learning Artificial Intelligence
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Quantization Meets Reasoning: Exploring LLM Low-Bit Quantization Degradation for Mathematical Reasoning
by: Li, Zhen, et al.
Published: (2025)

PTQTP: Post-Training Quantization to Trit-Planes for Large Language Models
by: Xiao, He, et al.
Published: (2025)

InfiR2: A Comprehensive FP8 Training Recipe for Reasoning-Enhanced Language Models
by: Wang, Wenjun, et al.
Published: (2025)

InfiAlign: A Scalable and Sample-Efficient Framework for Aligning LLMs to Enhance Reasoning Capabilities
by: Cai, Shuo, et al.
Published: (2025)

InfiMed: Low-Resource Medical MLLMs with Advancing Understanding and Reasoning
by: Liu, Zeyu, et al.
Published: (2025)

Revisiting Model Interpolation for Efficient Reasoning
by: Wu, Taiqiang, et al.
Published: (2025)

Exploring Layer-wise Information Effectiveness for Post-Training Quantization in Small Language Models
by: Xiao, He, et al.
Published: (2025)

Can We Trust LLMs on Memristors? Diving into Reasoning Ability under Non-Ideality
by: Wu, Taiqiang, et al.
Published: (2026)

BPDQ: Bit-Plane Decomposition Quantization on a Variable Grid for Large Language Models
by: Chen, Junyu, et al.
Published: (2026)

InfiAgent: An Infinite-Horizon Framework for General-Purpose Autonomous Agents
by: Yu, Chenglin, et al.
Published: (2026)

InfiGUI-R1: Advancing Multimodal GUI Agents from Reactive Actors to Deliberative Reasoners
by: Liu, Yuhang, et al.
Published: (2025)

LogicPuzzleRL: Cultivating Robust Mathematical Reasoning in LLMs via Reinforcement Learning
by: Wong, Zhen Hao, et al.
Published: (2025)

InfiMed-ORBIT: Aligning LLMs on Open-Ended Complex Tasks via Rubric-Based Incremental Training
by: Wang, Pengkai, et al.
Published: (2025)

Benchmarking LLMs' Mathematical Reasoning with Unseen Random Variables Questions
by: Hong, Zijin, et al.
Published: (2025)

InfiGUIAgent: A Multimodal Generalist GUI Agent with Native Reasoning and Reflection
by: Liu, Yuhang, et al.
Published: (2025)

Binary Weight Multi-Bit Activation Quantization for Compute-in-Memory CNN Accelerators
by: Zhou, Wenyong, et al.
Published: (2025)

Infi-MMR: Curriculum-based Unlocking Multimodal Reasoning via Phased Reinforcement Learning in Multimodal Small Language Models
by: Liu, Zeyu, et al.
Published: (2025)

InfiAgent: Self-Evolving Pyramid Agent Framework for Infinite Scenarios
by: Yu, Chenglin, et al.
Published: (2025)

MuonQ: Enhancing Low-Bit Muon Quantization via Directional Fidelity Optimization
by: Su, Yupeng, et al.
Published: (2026)

What Makes Low-Bit Quantization-Aware Training Work for Reasoning LLMs? A Systematic Study
by: Lv, Keyu, et al.
Published: (2026)

LFQ: Logit-aware Final-block Quantization for Boosting the Generation Quality of Low-Bit Quantized LLMs
by: Lee, Jung Hyun, et al.
Published: (2026)

APTQ: Attention-aware Post-Training Mixed-Precision Quantization for Large Language Models
by: Guan, Ziyi, et al.
Published: (2024)

InfiMed-Foundation: Pioneering Advanced Multimodal Medical Models with Compute-Efficient Pre-Training and Multi-Stage Fine-Tuning
by: Zhu, Guanghao, et al.
Published: (2025)

Atomic Thinking of LLMs: Decoupling and Exploring Mathematical Reasoning Abilities
by: Kuang, Jiayi, et al.
Published: (2025)

Access Sets Matter: Budgeting Expert Reads for Scalable Weight-Space Model Merging
by: Wang, Yuanyi, et al.
Published: (2026)

DyBit: Dynamic Bit-Precision Numbers for Efficient Quantized Neural Network Inference
by: Zhou, Jiajun, et al.
Published: (2023)

Shadow-FT: Tuning Instruct Model via Training on Paired Base Model
by: Wu, Taiqiang, et al.
Published: (2025)

InfiR : Crafting Effective Small Language Models and Multimodal Small Language Models in Reasoning
by: Xie, Congkai, et al.
Published: (2025)

Timber: Training-free Instruct Model Refining with Base via Effective Rank
by: Wu, Taiqiang, et al.
Published: (2025)

Rethinking Kullback-Leibler Divergence in Knowledge Distillation for Large Language Models
by: Wu, Taiqiang, et al.
Published: (2024)

InfoQuant: Shaping Activation Distributions for Low-Bit LLM Quantization
by: Li, Ke, et al.
Published: (2026)

On-the-Fly Adaptation to Quantization: Configuration-Aware LoRA for Efficient Fine-Tuning of Quantized LLMs
by: Ye, Rongguang, et al.
Published: (2025)

DART: Difficulty-Adaptive Reasoning Truncation for Efficient Large Language Models
by: Zhang, Ruofan, et al.
Published: (2025)

BitCal-TTS: Bit-Calibrated Test-Time Scaling for Quantized Reasoning Models
by: Patarlapalli, Sai Babu, et al.
Published: (2026)

LongFaith: Enhancing Long-Context Reasoning in LLMs with Faithful Synthetic Data
by: Yang, Cehao, et al.
Published: (2025)

FineQ: Software-Hardware Co-Design for Low-Bit Fine-Grained Mixed-Precision Quantization of LLMs
by: Xie, Xilong, et al.
Published: (2025)

Legal Mathematical Reasoning with LLMs: Procedural Alignment through Two-Stage Reinforcement Learning
by: Zhang, Kepu, et al.
Published: (2025)

Bit-by-Bit: Progressive QAT Strategy with Outlier Channel Splitting for Stable Low-Bit LLMs
by: Xu, Binxing, et al.
Published: (2026)

Low-Bit Quantization Favors Undertrained LLMs: Scaling Laws for Quantized LLMs with 100T Training Tokens
by: Ouyang, Xu, et al.
Published: (2024)

MergePipe: A Budget-Aware Parameter Management System for Scalable LLM Merging
by: Wang, Yuanyi, et al.
Published: (2026)