hub README + gates (393adcc)

hub README + gates (393adcc)

hub README + gates (393adcc)

hub README + gates (393adcc)

hub README + gates (393adcc)

Upload scripts/run_humaneval_raw.py with huggingface_hub

Upload scripts/run_humaneval.py with huggingface_hub

Upload scripts/auto_fuse.py with huggingface_hub

Upload scripts/fuse_layers.py with huggingface_hub

Upload scripts/run_humaneval.py with huggingface_hub

raw-completion HE runner (broken chat templates bypass) + HE_LIMIT smokes

HUMANEVAL_SERVER_ARGS extra flags (chat-template-file)

auto_fuse rank-vote (manual era ends), M7 corrected map

run_humaneval: HUMANEVAL_SERVE_MODEL self-serve + always-kill (no ghost servers)

verify flags (gpqa/he/all) + MERNIK rename + tri-fusion tooling and ledger

legacy: scripts/test_ffn_down.py

legacy: scripts/sensitivity_sweep.py

legacy: scripts/eval_humaneval.py

code: layer fusion

code: graft_mtp3 hybrid grafter

harness: fixed GPQA scorer (first-token top-k)

harness: GPQA duel

harness: TDN builder

harness: audit flags --big --layer

harness: audit_tiers.py

harness: scripts/mc_duel.py

harness: scripts/run_humaneval.py

harness: scripts/group_damage_sweep.py