arXiv cs.AI· Jan Kirin·· 10 小时前AI 评分31
循环语言模型中的操作性原型内省:在答案之前读取正确性
Operational Proto-Introspection in Looped Language Models: Reading Correctness Before the Answer
AI 导读
研究者在 2.6B 参数的循环模型 Ouro-RLTT 上探测隐藏状态,仅读取推理前 48 个 token 即可在合成逻辑任务上达到 0.758 AUROC,而 token 置信度仅 0.515。同一问题的两个答案间,探针有 74% 的概率给正确答案更高分,但该效应在 GSM8K 和第二个 checkpoint 上未获验证。探针信号可改善弃答判断,却无法让模型生成更好的答案。
正文 · 原文
Abstract:Can one tell, while a language model reasons, whether its answer will be right? Looped models are a good place to ask: they run the same layers several times per token, leaving one hidden state per loop for a probe to read. I probe Ouro-RLTT, a 2.6B-parameter looped model, for what these states reveal about how its reasoning will end. I call such information operational proto-introspection: operational because it is defined by what a probe can read, and proto because nothing here shows that the model itself uses it.
In part, yes. Reading the first 48 tokens of reasoning, from answers that state no answer yet, a probe on the hidden states reaches an AUROC of 0.758 on a synthetic logic task, against 0.515 for token confidence. Question difficulty is a rival explanation: on 1,164 further questions, token confidence plus simple features of the question predict about as well. It is not only difficulty: between two answers to the same question, the probe scores the correct one higher 74% of the time. A fine-tuned language model reading the same tokens does about as well. On GSM8K and on a second checkpoint this within-question test is unresolved.
A probe that ranks finished candidate answers is weak at layer 8 on the first loop, much better on the second, and as accurate as the last layers by the third, on four Ouro checkpoints. The direction it reads is largely shared across loops: a probe trained on the last loop reads the first nearly as well as a probe trained there.
The signal helps decide what to do with finished outputs. It improves abstention over surface features and picks correct answers among candidates better than chance, though surface features pick about as well. Once several samples are finished, a baseline that knows their agreement leaves the answer-free probe nothing measurable to add. None of my attempts to use it to make the model generate better answers worked.
| Comments: | Version 4 is substantially rewritten and adds fixed-prefix correctness tests, comparisons against visible-text readers and between answers to the same question, and cross-loop probe transfer. It revises the interpretation of the earlier final-answer-marker tests, corrects the ranking and pruning analyses, and removes unsupported preference claims. Appendix I details the changes |
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2607.18553 [cs.LG] |
| (or arXiv:2607.18553v4 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2607.18553 arXiv-issued DOI via DataCite |
Submission history
From: Jan Kirin [view email]
[v1]
Mon, 20 Jul 2026 22:40:36 UTC (375 KB)
[v2]
Sat, 25 Jul 2026 23:42:03 UTC (681 KB)
[v3]
Tue, 28 Jul 2026 16:28:19 UTC (864 KB)
[v4]
Wed, 18 Nov 2026 17:33:48 UTC (179 KB)
来源:arXiv cs.AI · arxiv.org