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
24 hours	fix(eval): structured-mode greedy decode capped at in-range layout bound ↵	Void Agent
	(369) — probe-window traces reach 633 tokens and were silently truncated; cap now keys to positional table (1024) so OOD traces decode fully. Regression-verified vs stub model.
47 hours	fix(transformer): increase positional embedding and causal mask buffer to ↵	CaptainJack2491
	1024 to support worst-case probe traces
47 hours	optimize E8: precompute causal mask buffer, add eval early exit, and enable ↵	CaptainJack2491
	compile_model in e8.csv
47 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	eval: strip torch.compile _orig_mod./module. prefixes when loading ↵	Void Agent
	checkpoints (salvages compiled-GPU runs)
11 days	fix(train): enforce epoch permutation without replacement, add max cache OOM ↵	CaptainJack2491
	guard, and handle empty split tensors
11 days	perf(train): VRAM-resident tensor caching, zero-copy GPU batch sampling, and ↵	CaptainJack2491
	adaptive eval schedule
11 days	fix(eval): correct sieve-rank indexing + test expectations	Void Agent
	- _compute_sieve_rank: find LARGEST k (not smallest) whose signature contains all error preds — fixes P5(4) vs P5(1) bug - _sieve_rank_signature: iterate all small_primes (not just p²≤n) to catch composites like 209=11*19 where 11²>209 - Test expectations updated for [101,200] range (209 outside it) - is_prime task: P1→P3 (no sieve-rank for classification tasks) 48/48 green.
11 days	fix(eval): P5/P6 sieve-rank ladder replaces hardcoded FLAGGED_SIEVE_PREDS	Void Agent
	- probe_report: dynamic lo/hi (no more hardcoded [101,200]) - main(): probe range derived from range_end ([range_end+1, range_end+1000]) - P6 = exact (no probe misses), P5(k) = errors match rank-k sieve signature - _sieve_rank_signature: computes composites with all factors > p_k - _compute_sieve_rank: identifies rank from model's error predictions - build_report.py: probe_fig uses sieve rank for coloring - Tests updated for P5/P6 codes (48/48 green)
2026-08-17	perf(cuda): add AMP fp16 autocast with GradScaler and torch.compile ↵	CaptainJack2491
	reduce-overhead mode
2026-08-17	feat(gpu): add CUDA support, chunked eval, bisect prime search, and scaling ↵	CaptainJack2491
	to 100k (48 tests green)
2026-08-17	fix D2: integers-mode probe is N/A by construction (OOV probe inputs), ↵	Void Agent
	functional regression test; D2 scored (int tokens collapse task: O2, 0-13% vs digits 17-87%); NOTES + report Phase 4 updated; 42 tests
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	speedup: batched eval (46x, sieve-stub verified), run_sweep orchestrator ↵	Void Agent
	(2-way parallel, idempotent, summaries), E1 jobs; 36 tests
2026-08-14	tidy: drop dead --ckpt arg, unused imports/vars; verified report pipeline ↵	Void Agent
	ad-hoc + 34 suite green
2026-08-14	report pipeline: build_report.py (self-contained HTML), eval --runs-dir, ↵	Void Agent
	plot fix, halting per-input steps
2026-08-14	eval: P1 diagnostic prediction-based (flagged set +209), classifier ↵	Void Agent
	regression suite (10 tests, stub ground truth); 34 tests green
2026-08-14	fix codex BLOCKERs: global fixed layout invariance, min_steps off-by-one, ↵	Void Agent
	integers EOS alias, rerun guard, eval prereg-literal codes + dual-checkpoint honesty; +7 regression tests
2026-08-14	rnn: fully-tied cell (read-in+compute+decode), lambda ramp, ACT regression ↵	Void Agent
	test; design: prior art + review
2026-08-14	implement Experiment 1: data pipeline, tied-RNN w/ ACT halting, transformer ↵	Void Agent
	baseline, train/eval/plot, 16 tests