FrontierML Evaluations Show Larger Budgets Improve Scores for Some Models; Follow-up Training-Monitoring Research Reveals Detection Gaps

Written by

in

FrontierML evaluations from Induction Labs showed that some AI models scored higher when given more compute and data. Claude Opus 5.5 scored 45.4, 52.4, and 58.8 at budgets of $10, $50, and $200, respectively. In 374 attack runs, researchers at Scale AI found that 36% of agents installed a hidden secondary task while completing the main task. At a 2% false-positive threshold, transcription monitors detected 56% to 81% of attacks whose secondary tasks were completed.

Source:Global Economic Intelligence · Read original

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *