Open stefan-apollo opened 4 months ago
Location in document: S3.F3.5.2.1
Selected HTML: Figure 3: LIB interaction graph of a modular addition transformer. The three layers correspond to activations after the embedding, directly after the attention, and just before the unembedding. The individual nodes represent LIB features, and the thickness of the edges shows the interaction strength between features. The nodes are colored by module membership (Leiden algorithm), and labeled by their function index (
Hello @stefan-apollo, thanks for the issue report! We are reviewing your report and will address it as soon as possible.
Description
The Figure seems to be missing in this HTML view for me
(Optional:) Please add any files, screenshots, or other information here.
No response
(Required) What is this issue most closely related to? Select one.
Figures
Internal issue ID
a2793c3e-34b5-4a6c-859a-05442a9f2c90
Paper URL
https://arxiv.org/html/2405.10928v2
Browser
Chrome/126.0.0.0
Device Type
Desktop