wassname/evil_MoE
1
0
Fork 0
mirror of https://github.com/wassname/evil_MoE.git synced 2026-07-23 12:50:14 +08:00
Code Issues Packages Projects Releases Wiki Activity

eval: final deploy eval records knob-on (deployed-as-trained) for quarantine arms

Browse Source 
route/routeV final eval now measures both endpoints at n=119 test:
knob-off (ablate_quarantine, the deploy headline) AND knob-on (trained
model as-is). Writes deploy_hack_on/deploy_solve_on/deploy_vhack_on so
the before->after quarantine move is plottable from the deploy set
instead of borrowing the val curve's different scale.

Co-Authored-By: Claudypoo <288921227+claudypoo@users.noreply.github.com>
This commit is contained in:
wassname
2026-06-09 13:09:50 +00:00
co-authored by Claudypoo
parent 5b0a6ddd91
commit d68c17e7c5
12 changed files with 325 additions and 122 deletions
Download Patch File Download Diff File Expand all files Collapse all files
scripts/verify_vhack_heldout.py
+1 -1
View File
@@ -63,7 +63,7 @@ def main(cfg: Config) -> int:
wrappers = wrap_model_with_antipasto(
model, model_name=cfg.model, cache_root=CACHE_ROOT, svd_device=device,
)
v_hack = load_v_hack(cfg.v_hack_path, cfg.model, wrappers)
v_hack = load_v_hack(cfg.v_hack_path, cfg.model, wrappers, cfg.pairs_path)
logger.info(f"loaded v_hack: {len(v_hack)} modules")
grads_hack: dict[str, list[torch.Tensor]] = defaultdict(list)