lyra
@_lyraaaa_
ˈli.ɹə 🏳️⚧️⚢ · 25 · ableton enjoyer · mechinterp · base model appreciator · evals and data farming · lyraaaa_ on discord · ♡ @bubblemoder ♡ 🦔~ ♪❀🫧 ⁸¹ ¹² ⁴¹
가입 May 2021
995 팔로잉 중 21.9K 팬
if you have a qwen3.5 4b around, please try steering neurons L14 N8949 +1 and L13 N7956 -1
now you have evil qwen
recipe: dose × median(‖resid‖ leaving layer L, tokens 1:) / ‖W_down[:, n]‖` at the down_proj at every input
더 보기
