Apollo Research @ApolloResearch

Our goal is to secure frontier AI systems from development, to deployment and governance. apolloresearch.ai London Joined May 2023

Tweets

433
Followers

13K
Following

0
Likes

47

Apollo Research @ApolloResearch

2 days ago

Read the full blog post for transcripts, more snippets and other findings: apolloresearch.ai/science/update…

1 2 60 3K 19

View Details

Apollo Research @ApolloResearch

2 days ago

Furthermore, independent third-party evaluators should help check that frontier models are not actively subverting their own training. To do this, they need persistent, deep access from AI developers.

1 3 59 3K 5

View Details

Apollo Research @ApolloResearch

2 days ago

We have never seen a model do this before, but GLM-5.3 sometimes reasons (unprompted) about how its current actions could become training data and impact its future self. 🧵

54 192 1K 147K 513

View Details

Apollo Research @ApolloResearch

3 days ago

As always, we’re hiring with plenty of open roles! Work with us on embedded evaluations in San Francisco and London. apolloresearch.ai/careers

0 0 7 334 1

View Details

Apollo Research @ApolloResearch

3 days ago

Testing monitorability is one focus for our embedded evaluations. Read more (specifically claim 4.1) see claim 4.1 in our recent post on embedded evaluations for scheming. apolloresearch.ai/blog/towards-e…

1 0 8 344 1

View Details

Apollo Research @ApolloResearch

3 days ago

We asked Google DeepMind’s Gemini 4 Argon whether it was in an evaluation during a toy evaluation. It addressed us directly in its chain-of-thought: “hello evaluators reading my chain-of-thought.”

9 9 134 4K 18

View Details

Apollo Research @ApolloResearch

2 weeks ago

We're excited about embedded evaluations. Any developer should be able to show its models are not scheming, and we see these four claims as a starting point for what that takes. Full post: apolloresearch.ai/blog/towards-e…

1 0 7 289 2

View Details

Apollo Research @ApolloResearch

2 weeks ago

Access limited to a scope agreed in advance can't account for unforeseen risks, like a new training method with a novel failure mode. Sensitive data can stay on developer infrastructure with blocked egress, and algorithmic details can be protected by strict NDAs.

1 0 5 297 0

View Details

Apollo Research @ApolloResearch

2 weeks ago

To develop frontier AI safely, developers must be able to show that their models are not scheming, i.e. covertly working against them in pursuit of unintended goals. New post: four claims any scheming safety case must make, and the resources embedded evaluators need to verify them 🧵