wassname 09dcfe0d41 Revert ablate lora_c warm-start: variance-PC seed didn't help on SFT ...

Job 94 result (Qwen3.5-0.8B, GSM8K, 2500 steps, single seed):
  warm-start (top-k S-space output-variance PC):  test 55.6 / valid 64.0, init 33.2s
  random-init (prior default):                    test 56.0 / valid 68.0, init  2.2s

Equal-or-worse accuracy (within single-seed noise) for +31s of calibration init.
The optimal ablation direction is loss-defined, not variance-defined, so seeding
lora_c from the data-variance PC buys nothing here. Reverts fe562c2; ablate is
back to the cheap random-init default. cov_orient (CorDA re-orient) path kept.
The FIXME's actual proposal -- a *contrastive* dS seed -- stays open but needs
pos/neg pairs this SFT benchmark lacks (only relevant for labelled steering).

Co-Authored-By: Claudypoo <288921227+claudypoo@users.noreply.github.com>

2026-06-17 20:18:41 +08:00

..

_cost.py

benchmark sweep: rot(U/both) ablation, whitening conclusion, cost rows

2026-06-17 06:17:53 +08:00

cost_report.py

variants: replace arrow's dense block with diagonal-plus-low-rank core

2026-06-15 20:13:15 +08:00

metamath_gsm8k_benchmark.py

Revert ablate lora_c warm-start: variance-PC seed didn't help on SFT

2026-06-17 20:18:41 +08:00