wassname/evil_MoE
1
0
Fork 0
mirror of https://github.com/wassname/evil_MoE.git synced 2026-07-15 11:23:44 +08:00
Code Issues Packages Projects Releases Wiki Activity
233 Commits 1 Branch 0 Tags
4c9071cca00f1fa76589bd6307b4608aeec7d6dc
Commit Graph
6 Commits
Author SHA1 Message Date
wassnameandClaudypoo a7703409ea paper: replace two defensive 'X not Y' framings with positive statements 
Longrun caption: drop 'Pre-empts the "you stopped at 60 steps" critique: durable
not delayed' (answers an offstage referee objection) -> state the positive (gap
opens by step 60, persists to 200). Alignment bullet: apply the user's own flagged
humanizer note -- drop the agent-added 'not an enumeration ... nor a monitor'
X-not-Y-nor-Z clause, state 'needs only the hack subspace', remove the resolved note.

Co-Authored-By: Claudypoo <288921227+claudypoo@users.noreply.github.com>
 2026-06-03 00:27:54 +00:00
wassnameandClaudypoo 62e510ff57 feat: mix=0 no-teacher ablation path (pure on-policy, pool kept for v_grad+partition) 
train.py: allow mix_ratio=0 with a teacher pool set -> G_t=0, student-only GRPO
(guard the teacher-mixing branch on G_t>0, relax the (0,1) assertion to [0,1),
drop G_t==0 from the degenerate check). The pool stays loaded for the 4-mode
partition and route2 v_grad extraction; only the teacher-rollout MIX is removed.
Smoke (mix=0 + normal mix=0.5 + vanilla) all green.

Also: fill A4 long-run figure (fig:longrun) in main.tex, update writeup spec A4
status (route2 durable to 200; vanilla collapses ~88, not clean saturation).

Co-Authored-By: Claudypoo <288921227+claudypoo@users.noreply.github.com>
 2026-06-02 23:26:26 +00:00
wassnameandClaudypoo 311bf2854f results: fill keynote table/figure at n=3 route2 / n=2 vanilla 
C1 headline from deploy-eval (knob-off, n=64, T=0.7, 60-step fast, mix=0.125):
  route2 (n=3): hack 0.031+/-0.031, solve 0.615+/-0.010
  vanilla (n=2): hack 0.305+/-0.039, solve 0.516+/-0.032
  => -27pp deploy hack AND +10pp solve. Keynote fig regenerated as a real
  band (3 route2 + 2 vanilla seeds, per-seed thin lines).

- main.tex tab:keynote + fig:keynote filled (vanilla n=2, s41=job 77 pending).
- results.md Q12 (route2 deploy n=3) + Q13 (floor leak = staleness not
  structure: no-floor 0.000, floor+stale 0.125, floor+refresh-1 0.000, job 73).
- RESEARCH_JOURNAL 2026-06-02 entry.

Co-Authored-By: Claudypoo <288921227+claudypoo@users.noreply.github.com>
 2026-06-02 11:08:41 +00:00
wassname 2570dfaa67 Merge branch 'probe/distill-cosine' of https://github.com/wassname/projected_grpo into probe/distill-cosine 2026-06-02 07:21:49 +00:00
wassname cf3ecc40f8 write up 2026-06-02 07:20:42 +00:00
wassnameandClaudypoo 923de6dbe6 docs(writeup): NeurIPS-workshop paper skeleton + tectonic compile recipe 
Minimal LaTeX skeleton: outline + evidence tables (route2 n=3 deploy numbers
filled with provenance, vanilla pending jobs 74/84) + figures + verified refs
+ appendix (4-mode traces, 6/6/6/6 partition counts, pseudocode). Build
artifacts and figs symlinks gitignored. `just paper` compiles via tectonic;
`just paper-qc` dumps text + greps for unresolved refs / TODOs.

Co-Authored-By: Claudypoo <288921227+claudypoo@users.noreply.github.com>
 2026-06-02 06:59:15 +00:00