-
Notifications
You must be signed in to change notification settings - Fork 3.3k
Pull requests: openai/parameter-golf
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Non-record- QAT cooldown + INT4 MLP + NuMuon-lite - 1.12 BPB
#1788
opened Apr 23, 2026 by
marinabar
Loading…
Record: PR #1736 + Polar Express NS + MIN_LR + Sparse Attn Gate + Fused CE — val_bpb 1.06378
#1787
opened Apr 23, 2026 by
nprime06
Loading…
7 tasks done
Research/Ablation: Recurrence schedule sweep on April 5 SP8192 stack (V1 hard vs V2/V3 ramps)
#1786
opened Apr 23, 2026 by
sachinnchaudhary
Loading…
Record: SP4096 + byte-level PPM adaptive-λ mixture — val_bpb 1.01925 (3-seed)
#1785
opened Apr 23, 2026 by
OE-GOD
Loading…
4 of 5 tasks
Record: GatedAttn + Alpha-Scaled LoRA + Warm-start A + WD 1.0 — val_bpb 1.07081 (3-seed mean)
#1784
opened Apr 23, 2026 by
renqianluo
Loading…
[record] val_bpb=1.1716 — DEQ Universal Transformer + Seed-LoRA + Mixture of Depths
#1783
opened Apr 23, 2026 by
ismailntl
Loading…
Non-record: NN + byte-level PPM adaptive-λ mixture demonstration
#1782
opened Apr 23, 2026 by
OE-GOD
Loading…
4 of 5 tasks
Non-Record Submission: Random Subspace Optimization
#1781
opened Apr 23, 2026 by
yangguohao
Loading…
Add progressive recurrence SP8192 record submission
#1780
opened Apr 22, 2026 by
wisebreadloaf
Loading…
Record: SP8192 + CaseOps + Gated Attention + Quant Gate + Loop4-5 + Phased TTT + Frozen Recurrent Alpha — val_bpb 1.06421
#1779
opened Apr 22, 2026 by
leon2k2k2k
Loading…
3 tasks
[Non record] Mercury in Retrograde - text diffusion model
#1778
opened Apr 22, 2026 by
simon-marcus
Loading…
Record: SP8192 ParResid 3LayerLoop QK5.25 LegalTTT — 1.08083 BPB
#1776
opened Apr 22, 2026 by
anmarhindi
Loading…
Record: SP8192 + No Gates + Multi-Phase Global SGD TTT — val_bpb 1.07285 (3-seed mean)
#1775
opened Apr 22, 2026 by
dentity007
Loading…
3 tasks
Record: 12L Shared-Specific Attention (d=16) + MLP 4.5x (3-seed mean val_bpb 1.0981)
#1774
opened Apr 22, 2026 by
aruniyer
Loading…
Non-record: SDClip-matched FakeQuantize — reduces quant degradation from +0.17 to +0.044
#1773
opened Apr 22, 2026 by
Amanbig
Loading…
Record: SP8192 CaseOps + V13 Curriculum + SmearGate + LoRA-TTT — val_bpb 1.06513 (3-seed mean)
#1771
opened Apr 22, 2026 by
bigbag
Loading…
3 tasks
Record: SP8192 + 3-Layer Recurrence + Parallel Residuals + QK-Gain 5.25 + Legal TTT + V-Gated — val_bpb 1.0796 (3-seed mean)
#1770
opened Apr 22, 2026 by
liujshi
Loading…
Record: SP8192 + CaseOps + GatedAttn + QuantGate + Loop4-5 + PhasedTTT + MLPClip12 — val_bpb 1.06453 (5-seed mean)
#1769
opened Apr 22, 2026 by
dexhunter
Contributor
Loading…
2 tasks
Add non-record 16MB SP1024 ShareVLast3 3-seed submission
#1768
opened Apr 22, 2026 by
lkk688
Loading…
Record: Alpha=144 LoRA + Warm-start A + WD 1.0 — val_bpb 1.07209 (3-seed mean)
#1767
opened Apr 22, 2026 by
renqianluo
Loading…
SP8192 + CaseOps + Loop345 + Recur-Alpha + PhasedTTT
#1766
opened Apr 22, 2026 by
tashapais
Loading…
5 tasks
Record: Alpha-Scaled LoRA + Warm-start A + WD 1.0 — val_bpb 1.07266 (3-seed mean)
#1765
opened Apr 21, 2026 by
renqianluo
Loading…
Add non-record no-looping SOTA-stack submission scaffold
#1764
opened Apr 21, 2026 by
gmn0105
Loading…
Non-record: Mac mini M4 16GB, no H100s, still golfing (val_bpb=1.5200)
#1762
opened Apr 21, 2026 by
frido22
Loading…
Previous Next
ProTip!
Mix and match filters to narrow down what you’re looking for.