giorgio/MachineSex
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

llm_society pilot v2: the contrast is alive — grounded society climbs, ungrounded decays

Browse source 
full: best 0.583 -> 0.617 and holds (a bred offspring beats every founder);
no_grounding: monotone decay to 0.533 despite identical elitism (conformist
selection keeps the wrong elite), with the confident-but-wrong gap growing to
+0.24. One seed - direction only; the 4-arm x 3-seed CX3 campaign is gated on
GG's review.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BkRLcc18rwT2Lysu6PbG7v
This commit is contained in:
Giorgio Gilestro 2026-09-07 13:49:51 +01:00
parent 70c8c54434
commit 9e83aa7034
1 changed files with 7 additions and 4 deletions
Download patch file Download diff file Expand all files Collapse all files

11
tasks/workorder-llm-society.md
View file

@ -74,10 +74,13 @@ disagreement), consensus accuracy, conformity−truth gap. E11's three panels re
first distillation generation and plateau at the same mutation–selection floor (best ≈0.40, mean first distillation generation and plateau at the same mutation–selection floor (best ≈0.40, mean
≈0.35, consensus 0.63→0.4 in both). Diagnosis: inheritance too lossy; the ~25%/generation ≈0.35, consensus 0.63→0.4 in both). Diagnosis: inheritance too lossy; the ~25%/generation
distillation tax swamps the selection differential. v1 archived at `results/llm_society_pilot1`. distillation tax swamps the selection differential. v1 archived at `results/llm_society_pilot1`.
- [ ] Pilot v2 (n_inherit=600, ep3, founders 600×3, elitism=1 all arms, n_parents=3, n_conf=90) — - [x] Pilot v2 (n_inherit=600, ep3, founders 600×3, elitism=1 all arms, n_parents=3, n_conf=90):
raises inheritance fidelity so selection has something to act on; elites = overlapping **the contrast is alive.** `full` climbs 0.583 → 0.617 best true fitness and holds (a bred
generations, applied identically in every arm (documented honestly; reproduction here is lossy offspring beat every founder at gen 3); `no_grounding` decays monotonically 0.583 → 0.533
distillation, unlike E11's faithful genotype copy) despite identical elitism — conformist selection keeps the wrong elite. Self-consumption
signature present: in `no_grounding` the population agrees with itself more (conformity 0.60 →
0.73) while the consensus gets less right (0.63 → 0.49); the confident-but-wrong gap reaches
+0.24 vs +0.16 in `full`. One seed, N=6: direction only. `results/llm_society`.
- [ ] GG gate: review pilot curves before the campaign - [ ] GG gate: review pilot curves before the campaign
- [ ] `hpc/llm_society.pbs` array job (arm × seed) → campaign - [ ] `hpc/llm_society.pbs` array job (arm × seed) → campaign
- [ ] Figure `paper/pnas/make_figs.py` panel(s); fold into main.md (replaces the "open" cell) - [ ] Figure `paper/pnas/make_figs.py` panel(s); fold into main.md (replaces the "open" cell)