AI Safety & Alignment
Jun 20, 2026
Resumo do dia
agenda: Interpretive debate. On “Model Organisms”. The distillation double bind: Distilling misaligned models either transfers misalignment or it doesn't
Principais notícias
- 1
agenda: Interpretive debate
agenda: Interpretive debate
- 2
On “Model Organisms”
On “Model Organisms”
- 3
The distillation double bind: Distilling misaligned models either transfers misalignment or it doesn't
The distillation double bind: Distilling misaligned models either transfers misalignment or it doesn't
- 4
Your Model Organisms Might Be Fried
Your Model Organisms Might Be Fried
- 5
Effective Altruism will be unbundled
Effective Altruism will be unbundled
- 6
Google DeepMind unveils plan to protect itself from its own rogue AI agents
Google DeepMind unveils plan to protect itself from its own rogue AI agents
Fontes
- agenda: Interpretive debate
- On “Model Organisms”
- The distillation double bind: Distilling misaligned models either transfers misalignment or it doesn't
- Your Model Organisms Might Be Fried
- Effective Altruism will be unbundled
- Google DeepMind unveils plan to protect itself from its own rogue AI agents
- Rackspace Technology (RXT) Is Up 42.5% After Cost-Cutting To Fund AMD-Powered AI Expansion
- Several frontier models are substantially prefill aware
- Alignement pretraining could backfire
- UK regulator sets out new rules on Google Search to boost competition
Share this with a friend
Send today's roundup to anyone who wants to keep up.
Get daily AI news free with AIToday
200+ AI sources, summarized in 1 minute. Email / LINE / Slack.
Sign up free