Research
Ai2 Releases StateBench to Measure Long-Context State Tracking
The Allen Institute for AI has released StateBench, a four-level benchmark designed to measure how well language models track changing variables across…
CodingResearch
#benchmark#long-context#evaluation#language-models#Ai2#AI#Benchmarks