claim
active
claim:automated-red-teaming-can-be-scaled-up-when-harmlessness-and-helpfulness-are-more-compatible-improving-robustnessAutomated red teaming can be scaled up when harmlessness and helpfulness are more compatible, improving robustness.
Section 6.1 suggests future work on scaling automated red teaming.
Source paper
extracted_from(2022) · Bai, Yuntao · Saurav Kadavath · Sandipan Kundu · Amanda Askell +47
Neighborhood — ranked by edge-count
Communities (2)
community
- CoT effects on generalization, multimodal QA accuracy, and AI safety alignment training.
- Comparative evaluation of RL-CAI and SL-CAI approaches for harmlessness using constitutional principles, 2022-2023 Anthropic research.
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- Out of 255 seeds x2.
- Interpretability insight accelerates red-teaming.
- Agent running many independent conversations to elicit concerning behaviors.
- Positions the work.
- Figure 2 and Figure 8 illustrate RL-CAI at the Pareto frontier.
- Stated generalization, untested here.
- Central thesis: expanding an agent's sensors and goals outward to include others' states creates bidirectional feedback loop that scales intelligence and increases compassion.
- Identifies recruitment as a cross-scale hallmark of collective intelligence.