Jim Lai

grimjim

·

AI & ML interests

Experimenting primarily with 7B-12B parameter text completion models. Not all models are intended for direct end use, but aim for research and/or educational purposes. Recent Contributions: stabilized refusal direction ablation via Gram-Schmidt orthonormalization and norm-preserving interventions; confirmed reasoning transfer via model merger.

Recent Activity

new activity 13 days ago

jukofyork/creative-writing-control-vectors-v3.0:“The doom lies in yourself, not in your name.”

updated a Space about 1 month ago

build-small-hackathon/dreadzone

published a Space about 1 month ago

grimjim/gradiochatbotexample

View all activity

Organizations

New activity in jukofyork/creative-writing-control-vectors-v3.0 13 days ago

“The doom lies in yourself, not in your name.”

#15 opened 11 months ago by

New activity in juiceb0xc0de/bella-bartender-gemma-e2b 2 months ago

subzero aikido flip

#1 opened 3 months ago by

New activity in grimjim/Daichi-Instructed-12B 4 months ago

Questioning utility.

#1 opened 4 months ago by

New activity in jwest33/qwen3-vl-8b-instruct-null-space-abliterated 5 months ago

Similar, parallel work?

#1 opened 5 months ago by

New activity in YanLabs/gemma-3-27b-it-abliterated-normpreserve 5 months ago

Something's broken

#1 opened 8 months ago by

New activity in grimjim/mistralai-Mistral-7B-v0.3 5 months ago

Query about the version

#1 opened 5 months ago by

New activity in MuXodious/OmniDimen-2-8B-Emotion-absolute-heresy 5 months ago

Fascinating model

#1 opened 5 months ago by

New activity in MuXodious/Goetia-24B-v1.3-absolute-heresy 5 months ago

Faithful decensor

#1 opened 5 months ago by

New activity in grimjim/gemma-3-12b-it-norm-preserved-biprojected-abliterated 7 months ago

add missing processor files

#3 opened 7 months ago by

New activity in grimjim/gemma-3-12b-it-biprojected-abliterated 7 months ago

What's the difference between projected and biprojected?

#1 opened 7 months ago by

New activity in grimjim/gemma-3-12b-it-MPOAdd-v1 7 months ago

Idea

#1 opened 7 months ago by

New activity in ArliAI/gpt-oss-20b-Derestricted 8 months ago

wtf man

#2 opened 8 months ago by

New activity in grimjim/gemma-3-12b-it-norm-preserved-biprojected-abliterated 8 months ago

Implementation Methodology

#2 opened 8 months ago by

New activity in grimjim/gemma-3-12b-it-projection-abliterated 9 months ago

Commendations

#2 opened 9 months ago by

New activity in anthracite-org/magnum-v4-12b 12 months ago

Recommended Parameters?

#10 opened 12 months ago by

New activity in grimjim/Llama-3-Luminurse-v0.2-OAS-8B-GGUF about 1 year ago

License Compatibility

#1 opened about 1 year ago by

New activity in grimjim/Magnolia-v5a-12B over 1 year ago

i didn't expect this to be that good 👌

#1 opened over 1 year ago by

New activity in open-llm-leaderboard/open_llm_leaderboard over 1 year ago

It's been a wild ride, folks :) (end of the Open LLM Leaderboard)

#1135 opened over 1 year ago by

New activity in grimjim/PAlign-PAPI-personality_prompt.json-cleaned over 1 year ago

Add task category and link to paper

#2 opened over 1 year ago by

New activity in open-llm-leaderboard/open_llm_leaderboard over 1 year ago

Spurious `trust_remote_code=True` objection when submitting a model?

#1100 opened over 1 year ago by