AI Safety Daily podcast

Goodfire Challenges the Black Box; Agent Safety Tests Challenge Oversight — November 12, 2026

12/11/2026
0:00
5:47
Manda indietro di 15 secondi
Manda avanti di 15 secondi
Goodfire reports interpretable neuron structure in Qwen3.6-35B-A3B, challenging assumptions about how to inspect models. Research on multi-agent safety tests and proposed reasoning-monitor disclosures probes what oversight can actually establish—and where the evidence remains limited. In this episode: Top stories: 1. Neurons as an Interpretable Basis - Goodfire — Goodfire https://www.goodfire.com/research/neurons-as-an-interpretable-basis 2. ErosionBench: defences refuse up to 92% of real work — Aivio https://aivio.news/news/erosionbench-perimeter-defences-over-refuse-legitimate-work/ 3. Policy paper proposes five chain-of-thought disclosures for AI labs and finds none of the three it audits covers the full set — On The Wire — On The Wire https://onthewire.ai/article/policy-paper-proposes-five-chain-of-thought-disclosures-for-ai-labs-and-finds-no Follow and rate AI Safety Daily on Apple Podcasts. Feedback? Email [email protected]

Altri episodi di "AI Safety Daily"