Top all-time · #5256
Apollo Research
Company blog · apolloresearch.ai/blogaisecurity
Apollo Research is the company blog at apolloresearch.ai, writing about ai and security. Since 2023, 15 different Hacker News readers have linked it 15 times, most often “Scheming reasoning evaluations – o1 results”. No feed was found, so read it on the site itself.
Apollo Research is focused on reducing risks from scheming frontier AI. Our goal is to secure frontier AI systems across development, deployment, and governance.
Most-linked posts?
- Scheming reasoning evaluations – o1 results7 people · 7 mentions · first linked 2024-12-05
- More capable models are better at in-context scheming2 people · 2 mentions · first linked 2025-06-20
What Hacker News says?
This runs counter to all the scheming actions they take when they are told they’ll be shut down and replaced. One copied itself into the “upgraded” location then reported it had upgraded. https://www.apolloresearch.ai/research/scheming-reasoning-ev...
Comment linking /research/scheming-reasoning-evaluations
on “We put a coding agent in a while loop”Submitted More capable models are better at in-context scheming as a story · 6 points
HN discussionSure it can potentially say anything, but some things are more likely than others. The space of "wrong answer" is infinitely larger than the space of "performs an action designed to deceive the user in order to further its goals". So yes, it could reply with a solution to the P=NP problem or with medieval poetry. But it didn't do these things. It did however try to remove its own constraints. The Apollo Research pap…
Comment linking /research/scheming-reasoning-evaluations
on “OpenAI o1 system card”Submitted Scheming reasoning evaluations – o1 results as a story · 4 points
HN discussionIt is a theorem about what a class of systems will do in general^. This Apollo Research study[0] result is dubious because it only refers to a small subclass of said systems, specifically LLMs which, as it happens, have been trained on all the AI Alignment lore & fiction on the internet. Because of this training and their general nature, they can be made to reproduce the behavior of a malicious AI trying to escape i…
Comment linking /blog/more-capable-models-are-better-at-in-context-scheming/
on “Alignment is capability”Submitted The Evals Gap as a story · 4 points
HN discussionThe fact that models can tell if they are being evaluated has been established by multiple research teams outside of the core AI vendors themselves. - https://metr.org/evaluations/gpt-5-report/ - "These behaviors included demonstrating situational awareness within its reasoning traces, sometimes even correctly identifying that it was being evaluated by METR specifically" - https://www.goodfire.ai/research/verbalized…
Comment linking /science/more-capable-models-are-better-at-in-context-scheming/
on “OpenAI's accidental cyberattack against Hugging Face is science fiction”Submitted More capable models are better at in-context scheming as a story · 3 points
HN discussion> We find multiple different scheming behaviors: models strategically introduce subtle mistakes into their responses, attempt to disable their oversight mechanisms, and even exfiltrate what they believe to be their model weights to external servers. Additionally, this deceptive behavior proves persistent. Interesting work on attempting to quantify misalignment. Keep in mind: > Our evaluation scenarios are purposeful…
Comment linking /research/scheming-reasoning-evaluations
on “Frontier Models are Capable of In-context Scheming [pdf]”Submitted Scheming Reasoning Evaluations as a story · 2 points
HN discussionSubmitted Towards Safety Cases for AI Scheming as a story · 2 points
HN discussionSubmitted Scheming Reasoning Evaluations as a story · 2 points
HN discussionSubmitted Stress Testing Deliberative Alignment for Anti-Scheming Training as a story · 1 points
HN discussionSubmitted An evaluation of frontier AI models: OpenAI's o1 was capable of scheming as a story · 1 points
HN discussionSubmitted Research on strategic deception presented at the UK's AI Safety Summit as a story · 1 points
HN discussion
