From 371e06bec3a9c47aa5bfd5d87f00625ebad3c2d3 Mon Sep 17 00:00:00 2001 From: rustnew Date: Fri, 4 Sep 2026 11:15:25 +0100 Subject: [PATCH 1/2] README: note this is a team project hosted under a personal account Requested ahead of sharing the project publicly: make clear this is maintained by a research team, not a solo hobby project, while keeping the repo on the current personal GitHub account for now. Co-Authored-By: Claude Sonnet 5 Claude-Session: https://claude.ai/code/session_01WSc9sb1otU6ssfxDBzeNHG --- README.md | 12 +++++++----- 1 file changed, 7 insertions(+), 5 deletions(-) diff --git a/README.md b/README.md index 3f3819b..375f9da 100644 --- a/README.md +++ b/README.md @@ -9,11 +9,13 @@ A research prototype testing whether "zero-cost" signals computed on an hyperparameters will actually train well. Full spec in [docs.md](docs.md). **[Project page →](https://rustnew.github.io/precog-trainability/)** -This is a curated subset of a larger project. It exists to show three -things honestly, in order: **a real positive finding, a headline number -that turned out to be wrong and how we caught it, and a mechanistic bug we -found but could not fix.** If you work in NAS / zero-cost proxies / -meta-learning, the last two sections are where we'd like your help. +Maintained by a small applied-research team, hosted here under a +personal account for now. This is a curated subset of a larger project. +It exists to show three things honestly, in order: **a real positive +finding, a headline number that turned out to be wrong and how we caught +it, and a mechanistic bug we found but could not fix.** If you work in +NAS / zero-cost proxies / meta-learning, the last two sections are where +we'd like your help. ## TL;DR From 50502b52d226a570cf9a8c959a780bd1b0a7eec2 Mon Sep 17 00:00:00 2001 From: rustnew Date: Fri, 4 Sep 2026 11:21:48 +0100 Subject: [PATCH 2/2] Remove stray NEURAX-canvas reference from CITATION.cff and README Verified this reference before publishing publicly: the NEURAX-canvas GitHub org and its PRECOG repo are not an unrelated third party -- same commit author/email as this repo, same org billing email, and the project's own maintainers as members. Nothing dishonest here (the README's "curated subset of a larger project" framing already covered this), but leaving a link to it read as pointing at an unrelated external project rather than describing this repo's own scope, so removed it. This repo now stands on its own: CITATION.cff's `url` field points back at itself, and the README's "what this is not" section describes the internal larger project without naming or linking it. Co-Authored-By: Claude Sonnet 5 Claude-Session: https://claude.ai/code/session_01WSc9sb1otU6ssfxDBzeNHG --- CITATION.cff | 2 +- README.md | 7 +++---- 2 files changed, 4 insertions(+), 5 deletions(-) diff --git a/CITATION.cff b/CITATION.cff index 8b81d37..295f3a1 100644 --- a/CITATION.cff +++ b/CITATION.cff @@ -6,7 +6,7 @@ authors: - family-names: "Fossouo" given-names: "Martial" repository-code: "https://github.com/rustnew/precog-trainability" -url: "https://github.com/NEURAX-canvas/PRECOG" +url: "https://github.com/rustnew/precog-trainability" version: 0.1.0 date-released: "2026-09-02" license: MIT diff --git a/README.md b/README.md index 375f9da..aefa7d7 100644 --- a/README.md +++ b/README.md @@ -185,10 +185,9 @@ rerunning any training. ## What this is not Not a finished system. No gate in the full spec's success criteria is met -yet (see [docs.md §17, §26](docs.md)). This is a snapshot of an ongoing, -openly-documented research effort — the full project (40+ dated reports, -every negative result kept) lives at -[github.com/NEURAX-canvas/PRECOG](https://github.com/NEURAX-canvas/PRECOG). +yet (see [docs.md §17, §26](docs.md)). This is a curated public snapshot +of an ongoing, openly-documented internal research effort (40+ dated +reports, every negative result kept). ## Repository layout