prime-grokking.git - Prime Grokking — can minimal architectures grok next-prime? Pre-registered negative-class experiment (weight-tied RNN vs transformer).

Age	Commit message (Collapse)	Author
25 hours	E8 launch order: armB (longest) first, then A/C/D1/D2/D3HEAD main	Void Agent

48 hours	optimize E8: precompute causal mask buffer, add eval early exit, and enable ↵	CaptainJack2491
	compile_model in e8.csv
48 hours	implement Addendum 8: 6-arm token-space recurrence suite (arms A, B, C, D1, ↵	CaptainJack2491
	D2, D3), tests, and jobs/e8.csv
7 days	E7 relaunch order: GPU cell first, completed tf-wd01 cell last (idempotent skip)	Void Agent

11 days	E7 jobs: 4M-step long-horizon, mixed-device (rnn->cpu, ↵	Void Agent
	transformer->cuda+compile)
2026-08-17	E6 jobs: 16 cells [2,1000] K=32 cuda AMP (14 primary + 2 batch-128 ↵	Void Agent
	ablation); Addendum 6 locked 61b002f
2026-08-16	phase3: task_mode (is_prime), train_frac, lr_decay flags + tests (41 green); ↵	Void Agent
	Addendum 5 locked; 11-job batch
2026-08-15	E2 wd sweep: 8 jobs (0.01/0.1/0.3/3.0 x both models, seed 0; 1.0 = existing ↵	Void Agent
	control)
2026-08-15	speedup: batched eval (46x, sieve-stub verified), run_sweep orchestrator ↵	Void Agent
	(2-way parallel, idempotent, summaries), E1 jobs; 36 tests