nils/jspace
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
d00f120f27142985a5dc93e45d4c4d22cabf5432
jspace/results-loop
T
History
NilsandClaude Fable 5 d00f120f27 PerDepthAdapter (Bae-style per-iteration merges) + convergence-halting probe (free ACT); pre-registration item 13 
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-15 01:07:01 +02:00
..
eval_bestof3_pi.json
best-of-3 oracle/deployable split: realistic selector loses the overall edge (55.0/27.3 vs oracle 57.8/34.5); paper economics rewritten
2026-07-14 17:16:00 +02:00
eval_bestof3_select.json
best-of-3 oracle/deployable split: realistic selector loses the overall edge (55.0/27.3 vs oracle 57.8/34.5); paper economics rewritten
2026-07-14 17:16:00 +02:00
eval_budgetcot_pi.json
budget-CoT per-item: paired ties with latent arms confirmed (p>=0.69)
2026-07-14 17:29:56 +02:00
eval_code_code_ff_full.json
n=500 controls: paired net-effect significant (loop vs untrained merge 17-4, p=0.007); ladder and net accounting finalized
2026-07-14 20:08:57 +02:00
eval_code_distill_s7.json
CPU endgame: stats pass (Wilson/McNemar/pooled hard), final figures, PAPER.md rewrite around amortizable-content thesis
2026-07-14 15:46:58 +02:00
eval_code_tap23_rerun.json
integrate LCB dissociation (untrained merge wins far-transfer, p=0.001 vs trained loop), capability panel (no MC damage), L23 exit completes sweep
2026-07-14 19:44:07 +02:00
eval_code_untrained_full.json
n=500 controls: paired net-effect significant (loop vs untrained merge 17-4, p=0.007); ladder and net accounting finalized
2026-07-14 20:08:57 +02:00
eval_lcb_distill.json
integrate LCB dissociation (untrained merge wins far-transfer, p=0.001 vs trained loop), capability panel (no MC damage), L23 exit completes sweep
2026-07-14 19:44:07 +02:00
eval_lcb_loop.json
integrate LCB dissociation (untrained merge wins far-transfer, p=0.001 vs trained loop), capability panel (no MC damage), L23 exit completes sweep
2026-07-14 19:44:07 +02:00
eval_lcb_untrained.json
integrate LCB dissociation (untrained merge wins far-transfer, p=0.001 vs trained loop), capability panel (no MC damage), L23 exit completes sweep
2026-07-14 19:44:07 +02:00
eval_panel_base_k0.json
integrate LCB dissociation (untrained merge wins far-transfer, p=0.001 vs trained loop), capability panel (no MC damage), L23 exit completes sweep
2026-07-14 19:44:07 +02:00
eval_panel_distill_ff.json
integrate LCB dissociation (untrained merge wins far-transfer, p=0.001 vs trained loop), capability panel (no MC damage), L23 exit completes sweep
2026-07-14 19:44:07 +02:00
eval_panel_loop_k2.json
integrate LCB dissociation (untrained merge wins far-transfer, p=0.001 vs trained loop), capability panel (no MC damage), L23 exit completes sweep
2026-07-14 19:44:07 +02:00
eval_rung2_s0.json
CPU endgame: stats pass (Wilson/McNemar/pooled hard), final figures, PAPER.md rewrite around amortizable-content thesis
2026-07-14 15:46:58 +02:00
eval_rung2_s1.json
CPU endgame: stats pass (Wilson/McNemar/pooled hard), final figures, PAPER.md rewrite around amortizable-content thesis
2026-07-14 15:46:58 +02:00
fig_ladder.png
n=500 controls: paired net-effect significant (loop vs untrained merge 17-4, p=0.007); ladder and net accounting finalized
2026-07-14 20:08:57 +02:00
fig_placement.png
CPU endgame: stats pass (Wilson/McNemar/pooled hard), final figures, PAPER.md rewrite around amortizable-content thesis
2026-07-14 15:46:58 +02:00
fig_scale.png
CPU endgame: stats pass (Wilson/McNemar/pooled hard), final figures, PAPER.md rewrite around amortizable-content thesis
2026-07-14 15:46:58 +02:00
fig_transfer.png
integrate LCB dissociation (untrained merge wins far-transfer, p=0.001 vs trained loop), capability panel (no MC damage), L23 exit completes sweep
2026-07-14 19:44:07 +02:00
lcb_labels.json
integrate LCB dissociation (untrained merge wins far-transfer, p=0.001 vs trained loop), capability panel (no MC damage), L23 exit completes sweep
2026-07-14 19:44:07 +02:00
PROTOCOL_UNIFIED.md
PerDepthAdapter (Bae-style per-iteration merges) + convergence-halting probe (free ACT); pre-registration item 13
2026-07-15 01:07:01 +02:00
REPORT_wave3.md
track protocol pre-registrations and interim report in git
2026-07-14 02:54:52 +02:00
stats_final.json
n=500 controls: paired net-effect significant (loop vs untrained merge 17-4, p=0.007); ladder and net accounting finalized
2026-07-14 20:08:57 +02:00
STATS.md
n=500 controls: paired net-effect significant (loop vs untrained merge 17-4, p=0.007); ladder and net accounting finalized
2026-07-14 20:08:57 +02:00