789 MB

Ctrl+K

1 contributor

History: 29 commits

AbstractPhil

README: explain the exp009->exp011 numbering (exp010 = seed program, see exp010_seeds/)

6270234 verified 18 days ago

exp001_relay
exp001_relay: the retrofit works at 0.5B (frozen 17.80 -> relay 14.09 / mlp 13.93, <0.6% trainable, 2 seeds); GPT-2 ppl ordering inverts (substrate-scoped) while the gate mechanism replicates (relay gates grow, MLP gates shrink) - the gate-vs-ppl dissociation 22 days ago
exp002_refine
exp002 Wave B seed-1 folded in: ordering budget-stable certified 2 seeds (mlp 13.899/13.889 vs relay_pw 13.983/13.932); relay convergence seed-dependent — claim corrected 22 days ago
exp003_instruct
exp003_instruct: hallucination reduction quantified (specialist grounding 0.979/0.958 vs frozen 0.542, >10x fewer failures); register capture found (specialist answers every register in its own); multi-task resolves it (all registers >=0.875); register-differentiation hypothesis CONFIRMED on the sign-code surface (2x deep-layer separation vs the wikitext stack, 2 seeds) 21 days ago
exp004_composite
exp004 amendment: seed-1 certification (truncated 0.0 both seeds, ok 1.0/0.938; over-emission replicates 1.667/1.311) — 2-seed rider satisfied 21 days ago
exp005_taskmoe
exp005 amendment: seed-1 rows — the s0 MATCH verdict is SEED-DEPENDENT (s1 nospec falls 0.917/0.708/0.563 while real composite holds 0.938/1.0/1.0): specialists = seed-robustness, echoing heredity-as-robustness 20 days ago
exp006_story
exp006_story: story anchors, 4 request registers — cont/inst/json 1.0, kw UNMOVED 0.083 (format learning != constraint compliance); strongest register sep at the time (L8 0.462) 21 days ago
exp007_math
exp007_math: math anchors, 4 ask-methods — format lock-in 4/4, no correctness tax, held-out f(x) 0.688->0.938; sep peak 0.496 @L23 21 days ago
exp008_tokendiff
exp008 amendment: 2-seed certification (denoiser replicates to 2nd decimal; sampler law both seeds; class-cond honest two-seed noise) 20 days ago
exp009_family
exp009 amendment: two-regime law s1 certification (story usage 0.821 vs 0.820; decouples replicate) + the preregistered OOD-band cell CONFIRMED (tokendiff usage 0.918 predicted >0.9; denoiser preserved to 3 decimals) 20 days ago
exp010_seeds
exp010_seeds index: the seed-ablation program (A1/A1b verdicts + amendment-row pointers) 18 days ago
exp011_newanchors
exp011_newanchors: infill anchor (+18pt over real LM floor, dose-response) + simesh similarity (infonce 0.494 recall@1 over 465 distractors; SIGN-CODE Hamming retrieval 0.359 — contrastive pressure reorganized the discrete surface; recon_target boundary marker 2:1) 20 days ago
exp012_bridges
exp012_bridges: CLIP bridge — first demonstrated cross-modal conditioning (0.273 vs 0.170 shuffled); BERT bridge honest negative x2 (predictability principle); roster consequence ledgered 20 days ago
exp013_collective
exp013_collective: 10-anchor collective — 3-for-3 preregistrations (infill 0.549 predicted 0.5-0.8; simesh passenger; tokendiff 0.886); full NxN matrix vs frozen row; task_1 starvation boundary + subcap attribution inconclusive (honest findings); mixed-batch per-position dispatch 20 days ago
exp014_toggles
exp014_toggles: the toggle registry live — gates bit-exact (all-off==frozen, strip==identity); span gating BEATS all-on 1.0 vs 0.875; inverted control 0.667; denominator-damping characterized; 3-segment registry demo ledgered 20 days ago
.gitattributes

1.52 kB
initial commit 22 days ago
README.md

12 kB
README: explain the exp009->exp011 numbering (exp010 = seed program, see exp010_seeds/) 18 days ago
article_ft2.md

31.6 kB
article_ft2 v2: full public citation pass — 29 external refs fetched+verified (0 hallucinated), 11 internal links verified, 3 corpus attributions corrected (Adam discipline -> memory-ft3; selector failure class -> ft1; attractors claim scoped), References + corpus sections added 21 days ago