-
Notifications
You must be signed in to change notification settings - Fork 1.5k
Pull requests: openai/parameter-golf
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Record: 12L Gradient-Guided Quant + Partial RoPE + LN Scale + EMA + XSA4 (val_bpb: 1.1320)
#332
opened Mar 21, 2026 by
saml212
Loading…
10L MLP3x + BigramHash(2048) + SWA + Stride-32: 1.1487 BPB
#331
opened Mar 21, 2026 by
Rhodrium
Loading…
3 tasks
Non-record: 11L Int6 + Online Logit Bias (val_bpb=1.1609)
#330
opened Mar 21, 2026 by
bopmite
Loading…
Non-record: MLX prototyping harness with validated technique stack (val_bpb=1.9588, Mac)
#328
opened Mar 21, 2026 by
kingjulio8238
Loading…
4 of 5 tasks
Submission TrigramHash + PartialRoPE + HeadTemp + stride32 (val_bpb: 1.1450)and
#327
opened Mar 21, 2026 by
Ananddna
Loading…
[Non-Record] QAT + NTK-4096 Eval + Cosine Warmdown + Aggressive SWA
#326
opened Mar 21, 2026 by
crony-io
Loading…
Add Looped Transformer Design non-record submission (non tuned)
#325
opened Mar 21, 2026 by
Aum08Desai
Loading…
Minimal recurrent motif (sb1 rs2 g0.18) – non-record submission
#323
opened Mar 21, 2026 by
megnat05-tmm
Loading…
11L SmearGate + BigramHash(10240) + Causal TTT + Mixed Int5/Int6 + SWA
#322
opened Mar 21, 2026 by
romainsantoli-web
•
Draft
Add record: Optimizer Tuning + Sliding Window Eval (val_bpb=1.1864)
#321
opened Mar 21, 2026 by
andreanjos
Loading…
Minimal recurrent motif (sb1 rs2 g0.18) – non-record submission
#320
opened Mar 21, 2026 by
megnat05-tmm
•
Draft
Non-record: Depth Recurrence 5x3 — Weight-Shared Looping Transformer (6xH200, val_bpb=1.2716)
#319
opened Mar 21, 2026 by
Arth-Singh
Loading…
5 of 8 tasks
Neural Cache: Cross-Window KV Caching for Extended Eval Context (research proposal)
#318
opened Mar 21, 2026 by
sseanliu
Loading…
3 tasks
Record: 11L XSA4 + EMA + TTT + Int6 MLP3x (val_bpb=1.1442)
#317
opened Mar 21, 2026 by
chris-buckley
Loading…
Non-record: 12L Low-Rank Q + QAT (1xH100, pre-quant 1.2035)
#316
opened Mar 21, 2026 by
SkywardSyntax
Loading…
3 of 6 tasks
Record: 11L Partial RoPE + LN Scale + EMA + Late QAT + XSA4 (val_bpb: 1.1248)
#315
opened Mar 21, 2026 by
jfprincz
Loading…
non-record: LR warmdown on 1x A40 (1.723 bpb, 8.40MB)
#313
opened Mar 21, 2026 by
my-sonicase
Loading…
Record: Int6 + Canon ACD (K=3) + Muon WD 0.04 + SWA + Sliding Eval (val_bpb=1.1668)
#312
opened Mar 21, 2026 by
chanwoo-park-official
Loading…
Non-record: Internal control port on the PR180 stack
#311
opened Mar 21, 2026 by
small-cactus
Loading…
Record: 10L Seq2048 TTT LoRA WarmdownQuant (val_bpb=1.1787)
#310
opened Mar 21, 2026 by
vishesh9131
Loading…
1 task done
Record: CLASE-Quant adaptive layer quantization (val_bpb=1.1914)
#309
opened Mar 21, 2026 by
NewyorkDev
Loading…
7 tasks done
Record: 11L XSA4 + EMA + Batch524K + zstd fallback (val_bpb: 1.1357)
#307
opened Mar 21, 2026 by
dennisimoo
Loading…
Previous Next
ProTip!
Updated in the last three days: updated:>2026-03-18.