3 results

capability-vector

inall models datasets apps agents people projects

Datasets

All datasets matching “capability-vector”

AlexWortega /capability-vectors capability-vectors — PASS / FAIL contrast trajectories This dataset bundles the exact PASS (comply) and FAIL (refuse) trajectories used to compute every direction in the AlexWortega/capability-vectors abliteration-experiments repo. Base model under study: AlexWortega/qwen35-4b-soyuz (LoRA on Qwen3.5-4B, Soyuz-sft). Each row is a single evaluation trajectory rendered to a chat-templated text blob, with the rollout's source bench, task_id, and either a numeric reward (comply) or a… See the full description on the dataset page: https://huggingface.co/datasets/AlexWortega/capability-vectors.textn<1K0 likes11 downloads4mo agoHugging Face