跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2607.04926 · v2

Input Pathways Shape Few-Shot, Not Zero-Shot, Binding in Tiny Transformers: A Fully-Enumerable Study

首次提交 2026-07-06 18:55
版本更新 2026-11-13 11:41
雷达观测 2026-11-17 04:57
arXiv ↗PDF ↗

摘要

How does the way information reaches a transformer -- as symbolic tokens, a clean per-factor "oracle" code, or an entangled perceptual vector -- shape whether it binds that information compositionally? We study ~6-10K-parameter transformers on finite factored worlds enumerated exhaustively, so every measurement covers the whole input space and the informative routes are information-matched (exact Bayes ceiling 1.0). Four findings. (1) Endpoint invariance: on held-out binding queries no informative route reaches converged zero-shot composition -- each ends at or below chance, so the failure reflects inductive bias under a lookup-sufficient objective, not missing information. It also holds with the state in its own sequence position rather than added to every token. (2) Readability governs few-shot binding: when a small fraction of the held-out type is leaked into training, sample efficiency tracks code readability. The least readable entangled code is worst (-0.147 pooled, under both schemes); a dimension-matched control and a graded readability sweep isolate readability from input dimension. (3) The injection site, not the pathway: v1 read the remaining ordering among readable routes as pathway parameter sharing; that is withdrawn. The compared pathways are block-decomposable, hence one function class, so sharing was never manipulated; the ordering comes from the initialization scale of the broadcast-added state and vanishes when it gets its own position (+0.057 -> -0.006, 20 paired seeds/cell). (4) Failure anatomy: symbolic routes lose the answer at the readout; index routes mis-bind -- the answer stays decodable, yet an input intervention shows the output tracks the wrong slot. The central positive claim is readability; the others bound what a pathway comparison can establish. Code, manifests, and per-seed logs are released.

作者备注:v2 withdraws the first few-shot factor: the compared pathways are one function class, so sharing was never manipulated, and the residual effect is the injected state's initialization scale, which vanishes under prefix injection (new appendix). Also withdraws the above-chance transient and the double dissociation (a mean of per-seed maxima). Readability and endpoint invariance stand. 18 pages

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存,登录后共享已有结果。

版本差异

本站首次收录为 v2,暂无更早版本可对照。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v2 · 2026-11-13 11:41 · 最新观测版本

Input Pathways Shape Few-Shot, Not Zero-Shot, Binding in Tiny Transformers: A Fully-Enumerable Study

How does the way information reaches a transformer -- as symbolic tokens, a clean per-factor "oracle" code, or an entangled perceptual vector -- shape whether it binds that information compositionally? We study ~6-10K-parameter transformers on finite factored worlds enumerated exhaustively, so every measurement covers the whole input space and the informative routes are information-matched (exact Bayes ceiling 1.0). Four findings. (1) Endpoint invariance: on held-out binding queries no informative route reaches converged zero-shot composition -- each ends at or below chance, so the failure reflects inductive bias under a lookup-sufficient objective, not missing information. It also holds with the state in its own sequence position rather than added to every token. (2) Readability governs few-shot binding: when a small fraction of the held-out type is leaked into training, sample efficiency tracks code readability. The least readable entangled code is worst (-0.147 pooled, under both schemes); a dimension-matched control and a graded readability sweep isolate readability from input dimension. (3) The injection site, not the pathway: v1 read the remaining ordering among readable routes as pathway parameter sharing; that is withdrawn. The compared pathways are block-decomposable, hence one function class, so sharing was never manipulated; the ordering comes from the initialization scale of the broadcast-added state and vanishes when it gets its own position (+0.057 -> -0.006, 20 paired seeds/cell). (4) Failure anatomy: symbolic routes lose the answer at the readout; index routes mis-bind -- the answer stays decodable, yet an input intervention shows the output tracks the wrong slot. The central positive claim is readability; the others bound what a pathway comparison can establish. Code, manifests, and per-seed logs are released.

查看此版本解读 →查看此版本原文 ↗