FIDE ranks experimental designs by reference-relative law revision
It completes aggregate constraints against a frozen baseline and prefers the design needing the least dynamical correction.
3 stories, newest first · one card per story
It completes aggregate constraints against a frozen baseline and prefers the design needing the least dynamical correction.
Information-gain policies beat training- and inference-time BED baselines on Twenty Questions and transfer to Medi-Q.
WhatWorkedBench tests whether AI agents can predict experiment outcomes across 36 tasks.
Today 1.6K readers · 2.4K views · 776 reads · all time 189K views