5 papers
Humanity's Last Exam
Long Phan, Alice Gatti, Ziwen Han +1144
Benchmarks are important tools for tracking the rapid advancements in large language model (LLM) capabilities. However, benchmarks are not keeping pace in difficulty: LLMs now achi…
Forecasting AI Time Horizon Under Compute Slowdowns
Parker Whitfill, Ben Snodin, Joel Becker
METR's time horizon metric has grown exponentially since 2019, along with compute. However, it is unclear whether compute scaling will persist at current rates through 2030, raisin…
Note on Selection Bias in Observational Estimates of Algorithmic Progress
Parker Whitfill
Ho et. al (2024) attempts to estimate the degree of algorithmic progress from language models. They collect observational data on language models' loss and compute over time, and a…
Will Compute Bottlenecks Prevent an Intelligence Explosion?
Parker Whitfill, Cheryl Wu
The possibility of a rapid, "software-only" intelligence explosion brought on by AI's recursive self-improvement (RSI) is a subject of intense debate within the AI community. This…
Beyond Ordinal Preferences: Why Alignment Needs Cardinal Human Feedback
Parker Whitfill, Stewy Slocum
Alignment techniques for LLMs rely on optimizing preference-based objectives -- where these preferences are typically elicited as ordinal, binary choices between responses. Recent…