Skip to content

Record Submission: 0.9258 BPB — Kitchen Sink (7-gram + XSA6 + BigramHash4K + Cosine TTT)#776

Open
agalimova wants to merge 2 commits intoopenai:mainfrom
agalimova:submission/kitchen-sink-0.9258
Open

Record Submission: 0.9258 BPB — Kitchen Sink (7-gram + XSA6 + BigramHash4K + Cosine TTT)#776
agalimova wants to merge 2 commits intoopenai:mainfrom
agalimova:submission/kitchen-sink-0.9258

Conversation

@agalimova
Copy link
Copy Markdown

Summary

Changes from PR #741

Parameter Default Ours
XSA_LAST_N 4 6
BIGRAM_VOCAB_SIZE 2048 4096
NGRAM_ORDER 5 7
NGRAM_ALPHA_HIGH 0.40 0.50

Test plan

  • 2 seeds on 8xH100 SXM (torch 2.9+cu126, FA3)
  • Eval time under 10-min budget (~520s)
  • All runs under 16MB artifact limit
  • 3rd seed running (will update)

🤖 Generated with Claude Code

agalimova and others added 2 commits March 25, 2026 09:45
Built on PR openai#700 with hyperparameter improvements found via
autoresearch-multi combinatorial search:
- XSA_LAST_N=6 (extended from 4 to 6 layers)
- BIGRAM_VOCAB_SIZE=4096 (doubled from 2048)

3-seed mean: 1.1078 (std 0.0045)
Seeds: 42=1.1045, 1337=1.1061, 2025=1.1129

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…ash4K)

Built on PR openai#741 with hyperparameter improvements found via
autoresearch-multi combinatorial search:
- XSA_LAST_N=6, BIGRAM_VOCAB_SIZE=4096, NGRAM_ORDER=7, NGRAM_ALPHA_HIGH=0.50

2-seed mean: 0.9258 (seeds 1337=0.9249, 42=0.9266)
Eval time: ~520s (under 10-min budget)

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
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