-
Notifications
You must be signed in to change notification settings - Fork 2.6k
Pull requests: openai/parameter-golf
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
16L XSA-all + INT4 MLP QAT + GPTQ + EMA + Partial RoPE + 30ep Cosine TTT
#951
opened Mar 27, 2026 by
Bharath-970
Loading…
Two-Level Dirichlet Posterior + Phrase Cache — 0.11556 BPB (3-seed)
#948
opened Mar 27, 2026 by
dentity007
Loading…
6 tasks done
Non-record: Legal Neural-Only No-TTT Alt (8xH100) val_bpb=1.1576
#947
opened Mar 27, 2026 by
aamodbhatt
Loading…
Non-record: Legal Neural-Only No-TTT (8xH100) val_bpb=1.1606
#946
opened Mar 27, 2026 by
aamodbhatt
Loading…
Record: Order-16 Frozen N-gram Oracle + Learned Gate + TTT — val_bpb 0.0274 (3-seed mean)
#945
opened Mar 27, 2026 by
TimPietrusky
Loading…
6 tasks done
Record: Compliance-First Packed Causal Memory + Dirichlet Mixing — val_bpb 0.01654407 (3-seed mean)
#944
opened Mar 27, 2026 by
aamodbhatt
Loading…
9 tasks done
Record: Score-First TTT + Multi-Order N-gram Backoff (3-seed mean val_bpb=0.9581)
#940
opened Mar 27, 2026 by
antaloaalonso
Loading…
Non-record: GatedDeltaNet, 32K Context, Document-Boundary State Reset
#939
opened Mar 27, 2026 by
brian386
Loading…
[non-record] 1xH100 screening: compression + eval strategy
#938
opened Mar 27, 2026 by
numb3r33
Loading…
[Non-Record Submission] CompressedUT CE + EMA Export + Export-Aligned Late QAT (1.4457 BPB)
#937
opened Mar 27, 2026 by
mihir-s-05
Loading…
Add non-record submission: ReasonBorn-Tiny CPU Prototype 2026-03-27
#936
opened Mar 27, 2026 by
Electroiscoding
Loading…
WIP: 1.1xxx BPB - MDL-T Stack (LeakyReLU² + EMA + GPTQ-lite + LateQAT + warmdown3500 + int6+zstd-22)
#934
opened Mar 27, 2026 by
tuanaqeelbohoran
Loading…
1 of 5 tasks
Record: CacheMoney — 0.0804 BPB (3-seed mean, std 0.00003)
#933
opened Mar 27, 2026 by
haikosys
Loading…
Non-record: CoDA-GQA Differential Attention — First Differential Attention Submission (val_bpb=1.1580)
#932
opened Mar 27, 2026 by
anthony-maio
Loading…
Record: 0.0498 bpb - Packed Training N-gram Artifact + Learned Weighting Gate (updated)
#931
opened Mar 27, 2026 by
AnirudhRahul
Loading…
Non-record: Entropy-regularized QAT (WIP, pending compute)
#930
opened Mar 27, 2026 by
lamb356
Loading…
Add record: 9L MLP3x LeakyReLU(0.5)² QAT Int6+zstd (val_bpb=1.1653)
#929
opened Mar 27, 2026 by
andreanjos
Loading…
Non-record: XSA-all + mHC + Full QAT (val_bpb=1.1211)
#928
opened Mar 27, 2026 by
autocode-rayes
Loading…
Recursive Transformer 4B/7L + VE + QAT + TTT — val_bpb 1.1696 (3-seed mean)
#927
opened Mar 27, 2026 by
Tonyy1977
Loading…
Record: 11L EMA + GPTQ-lite + LeakyReLU^2 + QAT@0.15
#926
opened Mar 27, 2026 by
NandhuRajRK
Loading…
Record: Frozen N-gram Oracle (Order-16) + Score-First TTT (0.02807 BPB)
#925
opened Mar 27, 2026 by
THUQiXuan
Loading…
Order-16 Frozen N-gram Oracle + Score-First TTT (0.02801 BPB)
#924
opened Mar 27, 2026 by
THUQiXuan
Loading…
Previous Next
ProTip!
Follow long discussions with comments:>50.