Record submission: Poly5 Softcap + BigramHash(3072) + Wider GPTQ-lite…

[jimliu741523]

… + TempScale + Z-Loss

Builds on current SOTA (1.1194 BPB) with 6 targeted improvements:

1. Poly-5 softcap (better compile fusion than tanh, proven in ternary submission)
2. BigramHash(3072) (SOTA ablation: -0.0009 BPB from 2048→3072)
3. Wider GPTQ-lite percentile search (9 candidates vs 5, lower quant error)
4. Temperature scaling T=0.95 at eval (conservative sharpening)
5. Z-loss regularization 1e-4 (training stability, from PaLM/Gemini)
6. LZMA preset 9 compression (better ratio)

All SOTA techniques preserved: 11L 512d, LeakyReLU(0.5)², XSA4, Partial RoPE 16/64, LN Scale, EMA(0.997)+SWA, Legal TTT, Parallel Muon, GPTQ-lite int6.

[MatoTeziTanka]

Community Review — Record submission: Poly5 Softcap + BigramHash(3072) + Wider GPTQ-lite…

BPB: 1.1194 | Compliance: LOOKS CLEAN — score-first-per-chunk TTT (legal #1416/#1423 pattern)

What I found in the code (head SHA 7218f44101bf, file records/track_10min_16mb/2026-03-26_Poly5Softcap_BigramHash3072_WiderGPTQ_TempScale/train_gpt.py):

The TTT path at line 1095 implements the score-first-per-chunk pattern: each chunk is scored under torch.no_grad() / inference_mode() before the base_model.train() + SGD adaptation runs on that same chunk, with an is_last_chunk guard so the final chunk gets no adaptation pass. This is the structural shape the legal frontier uses (PRs #1416 erichroepke, #1423 aryanbhosale).

Per Issue #402 and Issue #677, TTT is legal when each token is scored before the adapter updates on it, and that's what the code does here — chunk ci is scored under weights adapted only on chunks 0..ci-1. No prequant_ttt_adapt_adamw(val_tokens, ...) multi-epoch fine-tune, no scored-region SLOT, no target-in-key n-gram cache.

CPU smoke test (CT2038 proteus-engine, 2026-04-11): import OK in 0.04s, dim=512, layers=11, vocab=1024, code=90877 B, SMOKE_TEST_PASS

Verdict: LOOKS CLEAN.

Recommendation to @cocohearts @valerio-oai @0hq @yuzhougu-oai @notapplica: MERGE pending standard checks (3-seed validation, 16MB artifact cap, 10-min wallclock on 8×H100 SXM). The compliance picture matches the legal reference frontier and no flags were raised by the classification pass.

Auto-classification caveat: this review was drafted by the AST-based classifier against a template derived from manually-reviewed cluster PRs (#1420, #1450, #1487, #1541, #1529, #1533, #1518). If I've misread a subtlety in your eval path — e.g., multi-epoch TTT that I mistook for single-pass, or a target-in-key lookup I missed in a helper function — please flag it and I'll re-run the audit manually.

---

Reviewed by @MatoTeziTanka — The Agora. CPU smoke test (CT2038 proteus-engine, 2026-04-11): import OK in 0.04s, dim=512, layers=11, vocab=1024, code=90877 B, SMOKE_TEST_PASS. Classification via deterministic AST-based classify_prs.py (pattern bank derived from ~65 manually-reviewed PRs earlier in the 2026-04-11 sweep). This review was auto-drafted from a template and spot-checked before posting — if the template misread your code, please call it out so I can iterate the classifier.