Skip to content

Pull requests: openai/parameter-golf

New pull request New
1,056 Open 501 Closed
1,056 Open 501 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

Notable Non-Record: Switched Deep Supervision (first DS submission)
#1629 opened Apr 15, 2026 by channyzf6 Loading…
3 of 5 tasks
SP8192 Depth Recurrence + Parallel Residuals + TTT (1.1921 BPB)
#1628 opened Apr 15, 2026 by yu314-coder Loading…
3 tasks
Evolutionary NAS on only a 5 year old MacBook; within 10% of baseline
#1627 opened Apr 14, 2026 by mike-ferguson Loading…
2 tasks done
Record: VarLen Attention + Fused MLP + Multi-Phase Global SGD TTT — val_bpb 1.07193 (3-seed mean)
#1626 opened Apr 14, 2026 by dexhunter Contributor Loading…
4 tasks
1
[Non-record] E2E TTT at 27M scale — negative result (val_bpb 1.1104, SP1024)
#1625 opened Apr 14, 2026 by ChideraIbe123 Loading…
4 tasks done
SAFE_SUBMISSION: run036-safe016 (1.05850 BPB)
#1624 opened Apr 14, 2026 by joshkmartinez Loading…
Record submission: Distill+IntraLoop SP1024 9x512 (val_bpb=1.1942)
#1623 opened Apr 14, 2026 by divagr18 Loading…
BESE v5.3: Novel 288-token tokenizer (non-record 16MB)
#1621 opened Apr 14, 2026 by mrbese Loading…
Submit Lim Shiaw Yong: 1.66 BPB 12MB Squeeze Architecture
#1620 opened Apr 14, 2026 by shiawyonglim Loading…
Submission/sp8192 depthrecur adamwttt
#1619 opened Apr 14, 2026 by AVINASH0052 Loading…
9
ConvCompressed Residual Language Model
#1618 opened Apr 14, 2026 by Magumbian Loading…
1
[Non-Record] Single H100 16mb 1.21bpb
#1617 opened Apr 14, 2026 by adityasasidhar Loading…
5 tasks done
2
Attempt/qk gain 5.5 deeper recurrence
#1616 opened Apr 14, 2026 by Vickyrrrrrr Loading…
Non-record: systems-fusion investigation + H-Net M1 pilot
#1615 opened Apr 14, 2026 by diaslmb Loading…
Non-record: CUDA port of PR #1612 recipe (H100 pending)
#1614 opened Apr 14, 2026 by seekerPrice Loading…
2 of 5 tasks
Add CustomGPT model implementation in train.py
#1613 opened Apr 14, 2026 by GITHUB-SUBASHk Loading…
Non-record: MLX tuned hyperparameters — 1.5096 BPB local (H100 pending)
#1612 opened Apr 14, 2026 by seekerPrice Loading…
3 of 5 tasks
Record: VarLenAttn + PhasingTTT - val_bpb 1.0728 (3-seed mean)
#1610 opened Apr 14, 2026 by romeerp Loading…
SKC-600 ternary: split engram (packed-lite + eval) and harden build pipeline
#1609 opened Apr 14, 2026 by Akhilesh-Gogikar Loading…
5 tasks
Non-record: 1x H100 SXM5 Explorations
#1608 opened Apr 14, 2026 by User123331 Loading…
Non-record: Nemotron-H Mamba-3 Hybrid + First SSM Depth Recurrence (1.4765 BPB)
#1607 opened Apr 14, 2026 by inin-zou Loading…
3 tasks
Non-Record v2: 7L UNet + Int8 QAT + EMA + Long Train — 1.3969 BPB (DGX Spark)
#1606 opened Apr 13, 2026 by AlirezaAlampour Loading…
[Non-record] Experimentation Summary: Autopsy of 100+ Experiments — What Worked, What Didn’t, Mind Map for LLM Agents, etc.
#1602 opened Apr 13, 2026 by SPThole Loading…
[non-record] Sharpness-Aware Minimization (SAM) Inner Loop for Meta-TTT
#1601 opened Apr 13, 2026 by SPThole Loading…
Non-record submission: HELIX and HELIX MoR K7R2 U-Net (architecture report + finalized metadata)
#1600 opened Apr 13, 2026 by sayujshah Loading…
ProTip! Type g p on any issue or pull request to go back to the pull request listing page.