Report

Language models' claimed ability to learn to detect when their activations are read

A user asks how models could learn to do this and points to a thread.

2 Sources, 2h ago, first seen 2h ago

TLDR

A user claims language models can learn to detect when their activations are read. The post asks how and points to a thread.

Today's Rank

#12

Combined views

—

2 Sources, first seen 2h ago

likes

2 Sources

@charles_irlRT @Sauers_: Language models can learn to detect when you read their activations. How? 🧵2h

@raphaelmilliereI never considered that this might be possible even with training, but in hindsight it makes sense. The idea that it could in principle happen in the wild without being trained for is wild to think about1h

Today's Rank

#12

Combined views

—

2 Sources, first seen 2h ago

— likes

—

— comments

— saves

— reposts

Featured Source

SA Sauers@Sauers_10:50 PM · Oct 17, 2026
Language models can learn to detect when you read their activations. How? 🧵

— comments

— saves

— reposts

Sentiment

Positive——Negative

Summary

Not enough discussion yet.

No sentiment analysis available yet.

Sentiment

Positive——Negative

Summary

Not enough discussion yet.

No sentiment analysis available yet.

Featured Source

SA Sauers@Sauers_10:50 PM · Oct 17, 2026
Language models can learn to detect when you read their activations. How? 🧵

Sentiment

Positive——Negative

Summary

Not enough discussion yet.

No sentiment analysis available yet.

Sentiment

Positive——Negative

Summary

Not enough discussion yet.

No sentiment analysis available yet.

Today's Rank

#12

Today's Rank

#12

2 Sources

@charles_irlRT @Sauers_: Language models can learn to detect when you read their activations. How? 🧵2h

@raphaelmilliereI never considered that this might be possible even with training, but in hindsight it makes sense. The idea that it could in principle happen in the wild without being trained for is wild to think about1h