Signals 4 · free daily AI digest

$S^3$: Spectral Null-Space Swap Makes Reasoning Models Efficient

Paper recorded by Signals 4 on 2026-09-29 in cs.CL. Abstract reproduced from arXiv; link to the original below.

Published 2026-09-29 on arXiv · recorded by Signals 4 on 2026-09-30

Category: cs.CL · 自然语言处理 · first seen 2026-09-30

Abstract

LLMs trained with Chain-of-thought excel in reasoning capability, but often come with excessive token cost. We find that the core of reasoning capacity lies in the Thinking model's weight component within the null space of a projection defined by the corresponding Non-thinking model's dominant singular directions, and removing the subspace component can largely improve reasoning efficiency without

Read on arXiv →

#11 most recent of 292 cs.CL papers we have recorded · ↑ newer: BITEM at the NTCIR-19 R2C2 Task: Predicting Confidence from Agentic RA · ↓ older: On Trajectory-Aware Training for Masked Diffusion Language Models
Cite this page: $S^3$: Spectral Null-Space Swap Makes Reasoning Models Efficient: the #11 most recent of 292 cs.CL papers we have recorded (as of 2026-09-29). Source: Signals 4 (Signals API) — https://data.jiangzhang.ca/signals4/t/papers/s-3-spectral-null-space-swap-makes-reasoning-models-efficient.html
Free to quote with attribution to “Signals 4 (Signals API)”. Machine-readable: papers.json
Related: More cs.CL papers · arXiv signals · All papers · Today in AI
Get 4 AI signals a day by email — free.
Subscribe free → See all plans →
Get 4 AI signals a day by email — free
All models · All repos · By company · Daily editions