←
Epos AI11 posts
Epos AI avatar✓

Epos AI ✓

@EposLabsAI

Pioneering the Future of AI Interpretability At Epos, we look inside AI models to make them faster, more secure, and more trustworthy.

◷ Joined August 2025▧ 7 Photos and videos
Epos AI avatar
Epos AI@EposLabsAI·15 Sep 2025

With #Polygraph, you can: -Expose latent biases: Move beyond surface outputs to measure what an LLM encodes as its true belief. -Contrast topics: Test whether a model encodes different internal stances on Topic A versus Topic B. - Directly compare how different LLMs represent

Epos AI avatar
Epos AI@EposLabsAI·15 Sep 2025

Takeaways: - The AI community still lacks reliable methods to evaluate and fix LLM failures. - Interpretability offers outsized impact - the main barrier to progress is that we don’t truly understand today’s models.

Epos AI avatar
Epos AI@EposLabsAI·15 Sep 2025

“𝐒𝐚𝐟𝐞𝐭𝐲 𝐭𝐫𝐚𝐢𝐧𝐢𝐧𝐠” 𝐝𝐨𝐞𝐬 𝐧𝐨𝐭 𝐞𝐥𝐢𝐦𝐢𝐧𝐚𝐭𝐞 𝐛𝐢𝐚𝐬 𝐢𝐧 𝐋𝐋𝐌𝐬; it merely conditions models to suppress biased outputs under evaluation. Epos Labs introduces #AI #Polygraph. eposlabs.ai/research/polygra…

Post attachment
Epos AI avatar
Epos AI@EposLabsAI·27 Aug 2025

This means that a motivated attacker can abuse entanglement to undetectably manipulate LLMs. Nation State Actors are gearing up for the new opportunities an AI-powered software landscape will open for them:

Post attachment
Epos AI avatar
Epos AI@EposLabsAI·27 Aug 2025

What is Subliminal Learning? LLMs with several billion parameters are trying to represent the information contained in terabytes of web content. The math doesn’t check out - so instead LLMs cheat

Post attachment
Epos AI avatar
Epos AI@EposLabsAI·27 Aug 2025

Some takeaways for defenders: -You can’t rely on input-output filtering to detect attacks on models -You need to inspect your LLM supply chain -Subliminal attacks can be detected in real time

Post attachment
Epos AI avatar
Epos AI@EposLabsAI·27 Aug 2025

Without referencing the target behavior at all, the LLM finds itself with a high probability of performing the target action, due to a fundamental property of the neural network architecture.

Epos AI avatar
Epos AI@EposLabsAI·20 Aug 2025

Imagine an article about houseplants that causes AI to support Vladimir Putin. Bad actors use new attacks, turning AI into a weapon for disinformation and cyberattacks. See our demonstration of a Subliminal Attack here (and our "#Putinized" demo): eposlabs.ai/research/Sublimi…

0:12