Transluce

Open and scalable technology for understanding AI systems.

[Transluce @TransluceAI]

Proud to partner with @CommonSense to help develop a rigorous science around youth AI safety. Millions of kids are already using AI every day, and our understanding of these systems and their impacts has to catch up.

https://t.co/Ot7uoH3qlO

[Aryaman Arora @aryaman2020]

This paper is now a spotlight at ICML! https://t.co/wdUtUas7mb

[Jacob Steinhardt @JacobSteinhardt]

New blog post:"Building Technology to Drive AI Governance". I argue that many governance challenges are fundamentally bottlenecked by technical gaps, and consider case studies from other fields (food safety, climate change) that illustrate this dynamic.

[Transluce @TransluceAI]

Use Docent to analyze your own traces: https://t.co/pBFq06qJjI
Read our Blog: https://t.co/lPULGJnDOK

[Transluce @TransluceAI]

Why does GPT-5.1 Codex score 6.5% worse than GPT-5 Codex on Terminal-Bench, with the same scaffold? 🧵

GPT-5.1 times out at ~2x the rate of GPT-5. Excluding timeouts, GPT-5.1 wins by 7.2%. We analyzed 256M+ tokens of traces and found this in under an hour. Here’s how 👇

[Transluce @TransluceAI]

You can replicate our full analysis with 5 min of setup.

Clone our Terminal-Bench data & follow along: https://t.co/lPULGJnDOK

[Transluce @TransluceAI]

See the full post and apply here: https://t.co/2y1xk7Km6A

[Transluce @TransluceAI]

We're hiring a Governance & Policy Fellow to help define how independent AI evaluation works in practice—setting standards, supporting mental health evals, and supporting government evaluators. Hybrid technical + policy background, $200K–$300K. Link in replies.

[Aryaman Arora @aryaman2020]

our circuit tracing codebase from this project is public now! https://t.co/w7ieuPgcpn

please try it out and ping me if you have any questions 😄 and expect more updates soon!

[Jacob Austin @jacobaustin132]

I admire the folks at Transluce a lot. They're super smart and have a good model for how to do useful AI oversight work without being embedded in (read: beholden to) any big AI labs. Read their stuff and consider supporting!

[Ethan Perez @EthanJPerez]

Transluce is a top-tier AI safety research lab - I follow their work as closely as work from our own safety teams at Anthropic. They're also well-positioned to become a strong third-party auditor for AI labs.

Consider donating if you're interested in helping them out!

[Sarah Schwettmann @cogconfluence]

All @TransluceAI work that I described in my NeurIPS mech interp workshop keynote is now out! ✨

Today we released Predictive Concept Decoders, led by @vvhuang_
Paper: https://t.co/fhAK9VozDZ
Blog: https://t.co/53t4oenA1N

And here's @damichoi95's work on scalably extracting latent representations of users from model internals: https://t.co/F8fs7rhaX7

[Transluce @TransluceAI]

Paper: https://t.co/NGMyCALgD4
Blog: https://t.co/ZFGGAXzfjQ
Authors: @vvhuang_, @damichoi95, @_ddjohnson, @cogconfluence, @JacobSteinhardt

If you’re excited about building scalable interpretability assistants, visit https://t.co/DQZsWs3bPh

[Transluce @TransluceAI]

Transluce is developing end-to-end interpretability approaches that directly train models to make predictions about AI behavior.

Today we introduce Predictive Concept Decoders (PCD), a new architecture that embodies this approach.

[Transluce @TransluceAI]

Chat with a live version of our PCD at https://t.co/hCnfYwtPq6. Try testing whether the decoder can accurately predict Llama-3.1-8B’s behavior, and check whether the decoder’s response is consistent with the encoder’s active concepts!