Serveurperso's picture

6 1 3

Serveurperso PRO

Serveurperso

·

https://www.serveurperso.com

AI & ML interests

None yet

Recent Activity

reacted to grimjim's post with 👍 about 9 hours ago

Going forward, I will be adopting the term Magnitude-Preserving Orthogonal Ablation (MPOA) for my recent work in mitigating model damage from abliteration. The technique potentially unlocks reasoning capacity previously occupied with safety refusal processing. For details, start here: https://huggingface.co/blog/grimjim/norm-preserving-biprojected-abliteration Showcase results: https://huggingface.co/grimjim/gemma-3-12b-it-norm-preserved-biprojected-abliterated (outperforms base instruct on UGI Leaderboard NatInt) (The existing name, while technically accurate, was a bit of a mouthful.)

updated a model 17 days ago

Serveurperso/OuteTTS-Voices

new activity 19 days ago

MiniMaxAI/MiniMax-M2:Missing official chat_template / unclear initialization of <think> interleaving in MiniMax-M2 ?

View all activity

Organizations

New activity in MiniMaxAI/MiniMax-M2 19 days ago

Missing official chat_template / unclear initialization of <think> interleaving in MiniMax-M2 ?

#27 opened 19 days ago by

New activity in mistralai/Magistral-Small-2509-GGUF about 1 month ago

way worse then qwen3 coder or devstral for coding not recommended :(

#3 opened about 2 months ago by

[THINK] token not recognized in llama-cpp-python

#1 opened 2 months ago by

New activity in openai/README 4 months ago

Woah

#1 opened 4 months ago by

New activity in discord-community/WelcomeBot 9 months ago

test

#1 opened 9 months ago by

test

#2 opened 9 months ago by

commented a paper 9 months ago

Qwen2.5-VL Technical Report

Paper • 2502.13923 • Published Feb 19 • 210 •