Skip to content

Update plan.md: reset baseline to 1.0781 BPB, reprioritize directions#4

Closed
dhruvjatkar wants to merge 1 commit intomainfrom
worktree-agent-ae360f06
Closed

Update plan.md: reset baseline to 1.0781 BPB, reprioritize directions#4
dhruvjatkar wants to merge 1 commit intomainfrom
worktree-agent-ae360f06

Conversation

@dhruvjatkar
Copy link
Copy Markdown
Owner

Summary

  • Sets 1.0781 BPB (PR Record: 30ep Cosine TTT on LeakyReLU² stack (3-seed mean val_bpb=1.0781) openai/parameter-golf#672, unmerged) as the new target to beat. Merged SOTA remains 1.1194.
  • Reorders Top 8 research directions around the constraint that TTT is maxed at 30 epochs (590s/600s eval budget). All priorities are now orthogonal to TTT.
  • Collapses ~1000 lines of stale Round 0-3.9 session logs into a concise historical summary with key findings.
  • Removes resolved blockers (flash_attn, SSH hangs, local runtime).
  • Adds fresh Round 1 section tracking 5 currently submitted experiments.

Test plan

Generated with Claude Code

PR openai#672 maxes TTT at 30 epochs (590s/600s eval budget), so all future
improvements must be orthogonal to TTT. This update:
- Sets 1.0781 BPB (PR openai#672) as the new target to beat
- Reorders Top 8 directions: XSA-all confirmed at #1, Full GPTQ #2,
  SwiGLU #3, Muon-VS #4, aggressive quant #5, MASA openai#6,
  depth recurrence openai#7 with int6 risk warning, AdEMAMix openai#8
- Deprioritizes TTT-related directions already exploited by PR openai#672
- Collapses ~1000 lines of stale Round 0-3.9 session logs into a
  concise historical summary
- Removes resolved blockers (flash_attn, SSH hangs, local runtime)
- Adds fresh Round 1 section with 5 submitted experiments

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
dhruvjatkar pushed a commit that referenced this pull request Mar 25, 2026
PR openai#672 maxes TTT at 30 epochs (590s/600s eval budget), so all future
improvements must be orthogonal to TTT. This update:
- Sets 1.0781 BPB (PR openai#672) as the new target to beat
- Reorders Top 8 directions: XSA-all confirmed at #1, Full GPTQ #2,
  SwiGLU #3, Muon-VS #4, aggressive quant #5, MASA openai#6,
  depth recurrence openai#7 with int6 risk warning, AdEMAMix openai#8
- Deprioritizes TTT-related directions already exploited by PR openai#672
- Collapses ~1000 lines of stale Round 0-3.9 session logs into a
  concise historical summary
- Removes resolved blockers (flash_attn, SSH hangs, local runtime)
- Adds fresh Round 1 section with 5 submitted experiments

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
@dhruvjatkar
Copy link
Copy Markdown
Owner Author

Merged directly to main via cherry-pick

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant