-
Notifications
You must be signed in to change notification settings - Fork 1.9k
Pull requests: openai/parameter-golf
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Non-record: Add submission track_non_record_16mb/2026-03-23_DepthRecurrent_TTT
#495
opened Mar 23, 2026 by
SergiuDeveloper
Loading…
Non-record: Phase 1 Legal Score-First TTT + Meta-TTT (FOMAML) — awaiting compute
#494
opened Mar 23, 2026 by
george11642
Loading…
5 tasks
Record: 11L EMA + Int6 + XSA + LeakyReLU² + Partial RoPE (val_bpb: 1.1309)
#493
opened Mar 23, 2026 by
parinzee
Loading…
Record: 11L XSA4 + EMA + Partial RoPE + Rank-8 TTT Hooks (1.1591 bpb)
#492
opened Mar 23, 2026 by
Divyesh-Thirukonda
Loading…
Record: 11L Value Residual + Gated Attention + AdamW TTT (val_bpb=1.0891)
#490
opened Mar 23, 2026 by
amaljithkuttamath
Loading…
Record: 7L MLP3x + BigramHash + SmearGate + TTT 5ep (mean val_bpb=1.1327)
#489
opened Mar 23, 2026 by
sofiabod
Loading…
Record: 11L Int6 QAT + Warmdown (val_bpb=1.3267, 1xH100)
#488
opened Mar 23, 2026 by
pkim02
Loading…
3 tasks
Non-record: Value Residual (-0.015 BPB) + Gated Attention (-0.003 BPB) on 11L Production Stack
#487
opened Mar 23, 2026 by
anantdgoel
Loading…
Record: 11L TrigramHash + ValueResidual + GradQuant + AdamW TTT (mean val_bpb=1.1132, best 1.1101)
#486
opened Mar 23, 2026 by
ndokutovich
Loading…
Record: 10L CountInitBigram + XSA + PartialRoPE (val_bpb=1.1522)
#485
opened Mar 23, 2026 by
harsha-gouru
Loading…
Non-record: EBLS (Empirical Bayes Layer Sharing) — learned sharing patterns
#484
opened Mar 23, 2026 by
Robby955
Loading…
Track 10min_16mb: PR #287 family rerun at 585s wallclock (mean val_bpb=1.1346)
#483
opened Mar 23, 2026 by
tmustier
Loading…
Record: Cosine TTT scheduling with per-layer lr — mean val_bpb=1.0970 (3 seeds)
#481
opened Mar 23, 2026 by
mrdavtan
Loading…
Non-record: MoE exploration + multi-bit quantization analysis
#480
opened Mar 23, 2026 by
imyesung
Loading…
Record: 10L QAT + SwiGLU + BigramHash(10240) (val_bpb=TBD)
#479
opened Mar 23, 2026 by
chirag-bajaj
Loading…
New SOTA: 1.12676 BPB - 11L XSA-all(11) + GPTQ-lite + EMA + Late QAT
#478
opened Mar 23, 2026 by
gowtham0992
Loading…
[Non-record] ABRAM_CHIP v2 — HECR int16 ultra compact — 34 KB — 0.50 bpb
#475
opened Mar 22, 2026 by
abrahaw123-cell
Loading…
Non-record: 6-Technique Stack — Catalytic Residuals + Value Residual + Gated Attention + BigramHash(10240) + 12L (val_bpb=1.1690)
#474
opened Mar 22, 2026 by
joshuaswarren
Loading…
Record: Legal Score-First TTT + Parallel Muon — val_bpb 1.1220 (3-seed mean)
#473
opened Mar 22, 2026 by
abaybektursun
Loading…
Add submission: 11L Frontier Stack + Value Residual + Gated Attention
#471
opened Mar 22, 2026 by
yuvrajyadav17
Loading…
Non-record: Shared-weight transformer with extended warmdown (1.1454 val_bpb)
#470
opened Mar 22, 2026 by
leofeasby
Loading…
Previous Next
ProTip!
Exclude everything labeled
bug with -label:bug.