minus-square🇰 🌀 🇱 🇦 🇳 🇦 🇰 🇮 linkfedilinkEnglish40•edit-21 day agoBenchmark it how, exactly? Can it accurately give me Pokedex info for any pokemon I ask it about? Because even Fandom wikis can’t do that and those are edited, assumingly, by humans.
minus-squareScottlinkfedilinkEnglish28•1 day agoLooks like they had it play the game. https://techcrunch.com/2025/02/24/anthropic-used-pokemon-to-benchmark-its-newest-ai-model/
Benchmark it how, exactly? Can it accurately give me Pokedex info for any pokemon I ask it about? Because even Fandom wikis can’t do that and those are edited, assumingly, by humans.
Looks like they had it play the game.
https://techcrunch.com/2025/02/24/anthropic-used-pokemon-to-benchmark-its-newest-ai-model/