• Home
  • Technology
  • Gaming
  • Entertainment
  • World & Business
  • Science
  • Sports
  • AI
HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
AI
Announcement

Snorkel AI expands Open Benchmarks Grants tenfold to a $30 million commitment

Snorkel AI says it is launching a red team to test open benchmarks and a fellowship for independent contributors.

1 Source, 1h ago, first seen 1h ago

TLDR

Snorkel AI announced a tenfold expansion of its Open Benchmarks Grants to a $30 million commitment. It says the expansion will support more diverse, continuously updated open benchmarks, a red team to test them and a fellowship for contributors developing new evaluation methods. A later post about the expansion argues that independently developed benchmarks can help prevent “benchmaxxing,” while noting that private benchmarks and ongoing evaluations also matter.

Combined views

—

1 Source, first seen 1h ago

— likes— comments— saves— reposts

Combined views

—

1 Source, first seen 1h ago

— likes— comments— saves— reposts

Sentiment

Positive——Negative

Summary

Not enough discussion yet.

No sentiment analysis available yet.

Featured Source

Sentiment

Positive——Negative

Summary

Not enough discussion yet.

No sentiment analysis available yet.

1 Source

Alex Ratner@ajratnerIncredibly excited about our expansion of @SnorkelAI Open Benchmark Grants to $30M! Including calls for more areas, a red-teaming initiative, and a research fellowships program. Our view is that *more* benchmarks, built more robustly, continuously, diversely, and independently is what helps prevent overfitting (or "benchmaxxing"), and maintains the critical utility of public benchmarks for open evaluation and inspection of frontier AI. Of course: developing safe, effective AI requires private benchmarks, continuous human + automated evals, etc as well (we're proud of the work we do here to support many labs and enterprises). But open, independent visibility into frontier AI is critical- for the future of empirical AI research, for safety and alignment, for societal visibility into frontier AI, and for positive progress ahead - and we're excited to support!1h
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI

    1 Source

    Alex Ratner@ajratnerIncredibly excited about our expansion of @SnorkelAI Open Benchmark Grants to $30M! Including calls for more areas, a red-teaming initiative, and a research fellowships program. Our view is that *more* benchmarks, built more robustly, continuously, diversely, and independently is what helps prevent overfitting (or "benchmaxxing"), and maintains the critical utility of public benchmarks for open evaluation and inspection of frontier AI. Of course: developing safe, effective AI requires private benchmarks, continuous human + automated evals, etc as well (we're proud of the work we do here to support many labs and enterprises). But open, independent visibility into frontier AI is critical- for the future of empirical AI research, for safety and alignment, for societal visibility into frontier AI, and for positive progress ahead - and we're excited to support!1h
    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet