Signals 4 · free daily AI digest

Present but Hard to Read: Probing Latent Reasoning with Minimal Pairs

Paper recorded by Signals 4 on 2026-10-20 in cs.CL. Abstract reproduced from arXiv; link to the original below.

Published 2026-10-20 on arXiv · recorded by Signals 4 on 2026-10-21

Category: cs.CL · 自然语言处理 · first seen 2026-10-21

Abstract

Probing studies of latent reasoning often read a model's state where the answer begins and take a failed probe as evidence that a value is absent. We test this with a digit copied from the question, in the looped model Ouro-1.4B and the standard transformer Qwen2.5-1.5B. Comparing prompts that differ only in that digit shows it is present at the last prompt token, in a code shared across questions

Read on arXiv →

#195 most recent of 660 cs.CL papers we have recorded · ↑ newer: BitMTP: When Multi-Token Prediction Meets Low-Bit Large Language Model · ↓ older: OptionWorld: A Benchmark for Diagnosing Failures of LLMs as Long-Term
Cite this page: Present but Hard to Read: Probing Latent Reasoning with Minimal Pairs: the #195 most recent of 660 cs.CL papers we have recorded (as of 2026-10-20). Source: Signals 4 (Signals API) — https://data.jiangzhang.ca/signals4/t/papers/present-but-hard-to-read-probing-latent-reasoning-with-minimal-pairs.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