We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Evaluation reports should therefore document the system version, environment and interaction, task specification, elicitation procedure, success criteria, and assumptions used to relate observed behavior to safety-relevant risk.Disputed (Jul 2026)
2mo ago
Priorities include: evaluations that still work even when a model can determine if it is being tested and may therefore understate its harmful capabilities (“sandbagging”) or conceal misalignment (“alignment faking”), backed by secure access for inde...more Unverifiable source (Jul 7, 2026)
added 1mo ago
According to common benchmarks, open weight models are estimated to be 3 to 12 months behind the frontier (varying across capabilities). Developers of frontier closed models have recently designated these closed models as ‘High’ risk for cyber and bi...more AI Verified source (Jul 2026)
added 1mo ago
Safety for open-weight models whose capabilities are nearing frontier closed models requires unique tools for open-weight model safety and ecosystem monitoring. [...] Accelerated public and private investment in research is required to keep pace with...more AI Verified source (Jul 2026)
added 2mo ago