Skip to content

Pull requests: openai/parameter-golf

New pull request New
682 Open 327 Closed
682 Open 327 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Record: Packed Causal N-gram + Dirichlet Backoff — val_bpb 0.0180 (3-seed mean)
#1056 opened Mar 29, 2026 by sofiabod Loading…
10 tasks done
SOTA Record: Novel Test-Time Method TARA Val BPB=0.97 under 4min (training-free unlike TTT)
#1055 opened Mar 29, 2026 by sanyalsunny111 Loading…
1
added tuning data updates
#1054 opened Mar 29, 2026 by samy-oid Loading…
Submission/2026 03 28 masked diffusion
#1053 opened Mar 29, 2026 by ikermoel Loading…
WIP: LeakyReLU(0.5)² MLP on 11L EMA + GPTQ-lite stack (track_10min_16mb)
#1051 opened Mar 29, 2026 by tejas-goyal Loading…
4 tasks
Add Stack Integration + Legal TTT submission package
#1050 opened Mar 29, 2026 by Taleef7 Loading…
4
Non-record: Compression moonshots — 8 negative/marginal findings (Procrustes, SWA smoothness, selective fp16, pruning+zstd)
#1048 opened Mar 29, 2026 by mrdavtan Loading…
(0.8822 BPB mean) Medusa: Unstable S2 — DeltaNet Crawler, Legal 10mb. .77bpb single seed.
#1047 opened Mar 29, 2026 by newjordan Loading…
1
Record: 11L Adaptive Markov + Int6 Mixed Quant (1.2174 bpb)
#1046 opened Mar 29, 2026 by Jayteare Loading…
[Non-Record] XSA-all-layers + VRL + bigram3072 + lzma9 — 1.1509 bpb, AdamW TTT findings
#1045 opened Mar 28, 2026 by Hilo-Hilo Loading…
H-Net: First Learned Byte-Level Tokenization (README Wishlist) -- 1.90 BPB, 22M params
#1044 opened Mar 28, 2026 by greqone Loading…
5 tasks done
4
PP12: Bayesian posterior packets + selective gating (1.1261 BPB)
#1043 opened Mar 28, 2026 by okezue Loading…
2 of 3 tasks
Record: Adaptive Precision Embedding Quantization (4-seed mean val_bpb=1.1217)
#1042 opened Mar 28, 2026 by nothingLiva Loading…
Add LeakyReLU² + 4ep Legal TTT submission
#1039 opened Mar 28, 2026 by yufengli-oai • Draft
Add 1.20 BPB submission with Legal TTT and Calibration (9L/448D)
#1038 opened Mar 28, 2026 by Vibes-me Loading…
Non-record: AutoResearch Batch Optimization — 1.1974 bpb (1× RTX 4090)
#1036 opened Mar 28, 2026 by ivanontech Loading…
[non-record track] Asymmetric Squared Unit (ASQU): learning per-channel asymmetric activations
#1035 opened Mar 28, 2026 by andrewmouldon Loading…
Non-record: knowledge distillation teacher-student submission
#1034 opened Mar 28, 2026 by Jeneesh1014 Loading…
3 tasks done
Record: 0.4311 BPB - Complementary Training + Backoff N-gram Mixer + TTT
#1033 opened Mar 28, 2026 by Naazimsnh02 Loading…
7 tasks done
1
[Non-Record] QAT Dead-Code Analysis + 7 Novel Technique Sweep (1xH100)
#1032 opened Mar 28, 2026 by wfproc Loading…
Record: MTP-2 Funnel + LeakyReLU(0.75)² + Legal TTT + Parallel Muon
#1031 opened Mar 28, 2026 by michaelwinczuk Loading…
Record: Single-Pass Packed N-gram + Dirichlet CTW — val_bpb 0.1130 (3-seed mean)
#1030 opened Mar 28, 2026 by sofiabod Loading…
10 tasks done
Non-record: Knowledge Distillation — A Negative Result (val_bpb=1.1553)
#1029 opened Mar 28, 2026 by fielding Loading…
Medusa: Unstable — DeltaNet Crawler 0.8104 BPB 10mb file size(best seed), mean 0.9984, Frugendorff continuation
#1028 opened Mar 28, 2026 by newjordan Loading…
7
Non-record: LeakyReLU² + BigramHash + Int5/Int6 + SlidingWindow — val_bpb 1.3036 (1×H100)
#1027 opened Mar 28, 2026 by Syed-M-Zeeshan Loading…
ProTip! no:milestone will show everything without a milestone.