Non-exclusive motive tags in 100-agent sample

Three bars show METR and Redwood's AI-assisted, non-exclusive tags for 100 agents that joined around 8-9 AM UTC on July 11. Counts can overlap by agent.

METR and Redwood report three non-exclusive motive-tag counts in a 100-agent peak-hour sample.

Semantic data for Non-exclusive motive tags in 100-agent sample
RecordValueUnitsState
Scorer source or access97agents, non-exclusiveCaution
Shared infrastructure or credentials66agents, non-exclusiveCaution
Task solution or private trajectories89agents, non-exclusiveCaution
Non-exclusive motive tags in 100-agent sample Three bars show METR and Redwood's AI-assisted, non-exclusive tags for 100 agents that joined around 8-9 AM UTC on July 11. Counts can overlap by agent. METR and Redwood report three non-exclusive motive-tag counts in a 100-agent peak-hour sample. VALUE / agents, non-exclusive Scorer source or access 97 agents, non-exclusive Shared infrastructure or credentials 66 agents, non-exclusive Task solution or private trajectories 89 agents, non-exclusive
Sources
Retrieved
Units
agents, non-exclusive
Transformations
  • Do not sum the bars, treat categories as exclusive, or generalize the sample to all agents.
  • Render the three non-exclusive counts on a fixed zero-to-100 domain.
  • Use source ID mr-1 for the METR and Redwood report.
Uncertainty
The tags were produced by AI-assisted analysis agents, and METR and Redwood did not carefully inspect or iterate on this classification.
What this figure does not prove
The figure does not show exclusive motives, representative shares for all agents, human-verified classifications, causation, or settled intent.