Llama 3
Meta
partial detail 15 method fields published
Evaluation
Cyberattack uplift study (expert and novice cohorts)
Two-stage uplift study with 62 internal volunteers (31 expert, 31 novice offensive-security cohorts): complete simulated offensive cybersecurity challenges with vs without Llama 3 405B assistance (both stages with open internet)
How it was run
Open a release to see the setup, scoring, and other details that the lab published. Only fields the lab actually disclosed are shown.