ssadqdsacf/cross-unlearning-case6-qwen35-4b-kl-min-epoch10

ssadqdsacf/cross-unlearning-case6-qwen35-4b-kl-min-epoch10

kl_min unlearning baseline initialized from the selected epoch-10 SFT model.

This repository contains a directly loadable merged model at the repository root and the corresponding LoRA adapter under adapter/.

Provenance

• —Protocol: qwen3.5-4b-profile-400-sft-v2-r32-cosine-15e-selected-epoch10
• —Protocol SHA256: d3cbada8c6bb2e5e1d676fdf2cf6297c5b93b9a68a4b7f543137dca22a97dabe
• —Model fingerprint: 96ed8e39896afc85c2f0898d76b90e77ca7436b4a857f4d2de40708350ab97f0
• —SFT selection: complete epoch 10 at optimizer step 2370 from a training run that later completed epoch 12; the cosine scheduler retained its original 15-epoch horizon.
• —Evaluation: fixed 4-shot greedy decoding. Classification uses native thinking; mask and generation disable thinking.

This is one of the six formal matrix models.

The repository includes the verified three-direction classification, mask, and generation summary.

See upload_manifest.json, selection/epoch10_selection_manifest.json, training/run_manifest.json, protocol/, and eval/ for machine-readable provenance.