MMLU-Pro holds steady at 85.0, AIME 2025 slightly improves to 89.3, while GPQA-Diamond dips from 80.7 to 79.9. Coding and agent benchmarks tell a similar story, with Codeforces ratings rising from ...
"Hotline: Cybersecurity and Privacy" tackles the philosophical, moral, strategic, and organizational quandaries related to ...
To counter these psychological tactics, McGuire argued for a form of cognitive inoculation that would work much like a ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback