Labels

GPT-2 family

Llama family

Mistral family

OPT family

Qwen family

Analysis / probing / lenses

Cross-architecture / model loading

Interventions / generation modification

I/O / saving artifacts

Pipeline / Step framework

Plotting / visualization

Recording / activation extraction

Something isn't working

Cross-architecture or dependency-version compatibility issue

Pull requests that update a dependency file

Improvements or additions to documentation

This issue or pull request already exists

New feature or request

Pull requests that update GitHub Actions code

Good for newcomers

CPU only

Single CUDA GPU

Apple MPS

Multi-GPU (device_map=auto)

Author has opened or is preparing a PR

Extra attention is needed

High: blocks users from using a feature

Low: minor confusion or polish

Medium: feature is usable but takes guesswork

This doesn't seem right