andyx10/dit-loras-interpreting
andyx10/dit-loras-interpreting Experimenting with interpreting write vectors over 100 hidden-topic model organisms fromdiff-interpretation-tuning/loras implementation We use 'self_attn.o_proj andmlp.down_proj` for write vectors: two per block across 36 blocks, with a total of 72 write vectors per organism. Jacobian Lens from Neuronpedia neuronpedia/jacobian-lens (qwen3-4b/jlens/Salesforce-wikitext/Qwen3-4B_jacobian_lens.pt) layout test100/… See the full description on the dataset page: https://huggingface.co/datasets/andyx10/dit-loras-interpreting.
| Author / organization | andyx10 |
|---|---|
| License (as declared) | mit |
| Downloads | 1.4K (as reported by Hugging Face) |
| Likes | 0 |
| Created | 18 September 2026 |
| Last updated at source | 19 September 2026 |
Popularity, last 3 days
1,355 now (-375 since 2026-10-16). Counts as reported by the source, snapshotted daily by Silytics.
More from andyx10
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 18 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.