1 paper · 1 filter
Idan Davidovich, Debargha Ganguly, Vikash Singh +1
Leading benchmarks for formal theorem proving with large language models are small collections drawn from competition math, such as the IMO and Putnam, that poorly represent field-…