30 秒でわかる
The speakers discuss jailbreak robustness and long-tail failures in language models used as judges. The speaker argues that subtle failures may still matter and that human supervision can be warranted.
- 1RobustnessGreater robustness does not imply failures have disappeared.
- 2Long-tail failuresSubtle failures can still matter.
- 3SupervisionAutomated judging requires evaluation and may need human supervision.
ポイント
- Greater robustness does not imply failures have disappeared.
- Subtle failures can still matter.
- Automated judging requires evaluation and may need human supervision.
チャプターごとに整えた全文を読む
Limits of automated evaluation
The speaker discusses failure modes and oversight.
The interviewer asks how to interpret claims that a model is more robust to jailbreaks: do obvious failures remain, or have the remaining cases become more subtle?
The speaker says mitigation can make failures subtler, but subtle does not mean harmless. Language models used as judges can have unusual, long-tail failure cases, as can refusal mechanisms.
There is a tradeoff between compute spent on an automated judge and its accuracy. The speaker says failures can still warrant human supervision; the clip does not quantify an error rate or define a universally safe threshold.
同じチャンネルの記事
1:49:40EN
Cognitive Revolution "How AI Changes Everything"· 2026年10月28日ポッドキャスト · 4 人A Strategy for Preserving Options on the Path to Advanced AIゲスト Joel Predd, Ben Boudreaux, Matt Chessen · ホスト Nathan LabenzRAND researchers discuss uncertainty, security architecture and human agency with Nathan Labenz.
1:28:36ENCognitive Revolution "How AI Changes Everything"· 2026年10月24日ポッドキャスト · 5 人AI’s Pace: Interpretability, GPU Credit and the Physical Supply Chainゲスト Zvi Mowshowitz, Evan Conrad, Lonis Hamaili · ホスト Nathan Labenz, Prakash NarayananA weekly highlights episode weighs uncertain AI research trajectories against finance, infrastructure, security and regulation.
3:13:44ENCognitive Revolution "How AI Changes Everything"· 2026年10月22日ポッドキャスト · 4 人AI’s Physical Bottlenecks: Compute Markets, Chip Investments and Agent Trustゲスト Evan Conrad, Lonis Hamaili · ホスト Nathan Labenz, Prakash NarayananA full AI:AM discussion of security incidents, GPU financing, semiconductor supply and an agent-to-agent network.
2:40:58ENCognitive Revolution "How AI Changes Everything"· 2026年10月21日ポッドキャスト · 3 人Frontier AI, Interpretability and the Fight Over Its Paceゲスト Zvi Mowshowitz · ホスト Nathan Labenz, Prakash NarayananZvi Mowshowitz joins AI:AM to discuss neural networks, policy and the institutions facing rapid change.
1:25:56ENCognitive Revolution "How AI Changes Everything"· 2026年10月17日ポッドキャストClaude Changed the Password. The Steering Wheel Was Already Optionalホスト Nathan LabenzNathan Labenz and Prakash Narayanan spend a week on agents that can lock you out, models that dodge a feeling they deny having, and formal proofs that found Linux bugs.
2:51:22ENCognitive Revolution "How AI Changes Everything"· 2026年10月14日ポッドキャスト · 4 人AI Minds and 8 mm Memories: Research, Archives and Simulationゲスト Cameron Berg, Phil Maher · ホスト Nathan Labenz, Prakash NarayananCameron Berg discusses hidden model states; Phil Maher explains preserving home movies and a fly-brain demo.