corpus.blog
56,966 blogs · 4,417,066 posts Claim your blog
Record · corpus.blog/posts/01a08758-be6e-709c-be3a-13c78fa38ba2

Training a Transformer to Compose One Step Per Layer (and Proving It)

brendanlong.com · published 26 April 2026

A record is what was published and who pointed at it. The text of the post is not held here: read it at the source, or at its Wayback capture.

Read the post at brendanlong.com Wayback capture
0blogs citing
First cited
—
Most recent
—
Rank this month
not ranked
Links from this post
10
Words captured
2,286

Cited by

No blog we hold has cited this post yet.

Links from this post

Link
Host
https://arxiv.org/abs/2406.17557
arxiv.org
https://github.com/karpathy/nanogpt
github.com
https://arxiv.org/abs/2201.02177
arxiv.org
https://arxiv.org/abs/2206.04301
arxiv.org
https://en.wikipedia.org/wiki/Dihedral_group
en.wikipedia.org
https://en.wikipedia.org/wiki/Non-abelian_group
en.wikipedia.org
https://arxiv.org/abs/2405.15071
arxiv.org
https://arxiv.org/abs/1807.03819
arxiv.org
https://www.lesswrong.com/posts/AcKRB8wDpdaN6v6ru/interpreting-gpt-the-logit-lens
lesswrong.com
https://github.com/brendanlong/sequential-transformer-lens-experiment
github.com