paper-with-me

Papers

Scalable LLM Math Reasoning Acceleration with Low-rank Distillation

2025-05-08 · Harry Dong, Bilge Acun, Beidi Chen, Yuejie Chi

Due to long generations, large language model (LLM) math reasoning demands significant computational resources and time. While many existing efficient inference methods have been developed with excellent performance preservation on language tasks, they often severely degrade math performance. In this paper, we propose Caprese, a low-cost distillation method to recover lost capabilities from deploying efficient inference methods, focused primarily in feedforward blocks. With original weights unperturbed, roughly 1% of additional parameters, and only 20K synthetic training samples, we are able to recover much if not all of the math capabilities lost from efficient inference for thinking LLMs and without harm to language tasks for instruct LLMs. Moreover, Caprese slashes the number of active parameters (~2B cut for Gemma 2 9B and Llama 3.1 8B) and integrates cleanly into existing model layers to reduce latency (>11% reduction to generate 2048 tokens with Qwen 2.5 14B) while encouraging response brevity.

📄 PDF Abstract BibTeX arXiv:2505.07861

Code (0)

등록된 구현이 없습니다.

Tasks

Language ModelingLanguage ModellingLarge Language ModelMath

Methods 이 논문이 사용한 방법론

LLaMA LLaMA is a collection of foundation language models ranging from 7B to 65B parameters. It is based on the transformer architecture with various improvements that were…

Similar Papers 제목 키워드 기반

LoRi: Low-Rank Distillation for Implicit Reasoning

2026-06-03 · Ryan Solgi, Jiayi Tian, Zheng Zhang arxiv

Implicit chain-of-thought (iCoT) methods aim to internalize reasoning in large language models, but often underperform explicit CoT prompting. We empirically find that hidden-state reasoning trajectories exhibit low-rank…

Mathematical Reasoning

ReM-MoA: Reasoning Memory Sustains Mixture-of-Agents Scaling

2026-06-23 · Heng Ping, Arijit Bhattacharjee, Peiyu Zhang, Shixuan Li 외 arxiv

Mixture-of-Agents (MoA) architectures improve inference-time scaling by organizing multiple LLM agents into layered reasoning pipelines. However, existing MoA variants fail to sustain gains as depth increases, exhibiting…

Formal Logic

VOLD: Reasoning Transfer from LLMs to Vision-Language Models via On-Policy Distillation

2025-10-27 · Walid Bousselham, Hilde Kuehne, Cordelia Schmid arxiv

Training vision-language models (VLMs) for complex reasoning remains a challenging task, i.a. due to the scarcity of high-quality image-text reasoning data. Conversely, text-based reasoning resources are abundant and sca…

Reinforcement Learning

Anti-Self-Distillation for Reasoning RL via Pointwise Mutual Information

2026-05-12 · Guobin Shen, Xiang Cheng, Chenxiao Zhao, Lei Huang 외 arxiv

On-policy self-distillation, where a student is pulled toward a copy of itself conditioned on privileged context (e.g., a verified solution or feedback), offers a promising direction for advancing reasoning capability wi…

Distillation and Refinement of Reasoning in Small Language Models for Document Re-ranking

2025-04-04 · Chris Samarinas, Hamed Zamani

We present a novel approach for training small language models for reasoning-intensive document ranking that combines knowledge distillation with reinforcement learning optimization. While existing methods often rely on …

Document RankingInformation RetrievalKnowledge DistillationLanguage Modeling+4