Parallax

EU-based non-profit research lab. Building the science of belief and goal elicitation for white-box auditing of frontier AI agents.

[Javier Ferrando @javifer_96]

Excited to have joined @parallx_ai! Looking forward to tackling important problems with a talented team!

[Mario Giulianelli @glnmario]

Back in Cambridge where Parallax started, and very happy to be here with this brilliant team. Proud of what we’re building together and looking forward to sharing our work on the science and infra of AI auditing 🖤

[Parallax @parallx_ai]

Meet the Parallax team! 🔍

Our founding team is gathering for the first time in London and Cambridge this week. We will kickstart our research activities and shape our agenda for bringing white-box belief auditing methods in the hands of AI evaluators.

It’s lovely to finally have everyone in one place!

[Gabriele Sarti @gsarti_]

I'll be at @emnlpmeeting as a co-organizer of @BlackboxNLP and co-author of "Demistifying Entropy-based Selection for CoT Compression in LRMs" led by the brilliant @SaraCandussio & @daniel_sc4!

Ping me here or on Whova to chat about scalable interp for long-horizon agent auditing & research opportunities @parallx_ai 🔍

[Gabriele Sarti @gsarti_]

Aletheia's Quest winners are finally out! 🏆

This summer, our @ndif_team at @Northeastern teamed up with @cadenza_labs with the support of @schmidtsciences and @AWS to challenge AI researchers to build generalizable LLM lie detectors, with $50k in prizes.

After 478 submissions by 19 teams, final results are in 🧵 1/

[Gabriele Sarti @gsarti_]

Why not both? At @parallx_ai, we believe that infrastructure for double-blind audits like the one piloted by @AVERIorg, @openminedorg, @GoogleDeepMind and @MLCommons (https://t.co/boSs2VtDli) will play a fundamental role in the future of white-box audits. Still, model-specific adaptations would likely require embedded access to maximize effectiveness. A balance of both is probably the most reasonable way forward.

[David Bau @davidbau]

There is a gap between the telescope of blackbox eval and the microscope of whitebox interpretability. I'm excited by @gsarti_'s new European org that aims to bridge this gap.

@ndif_team will partner with them.

Online: https://t.co/P5q83phuXq (drop the a). They're hiring!

[Gabriele Sarti @gsarti_]

Excited to kickstart this effort with @glnmario and a star-studded advisory team! 🚀

Also, we're hiring for 2 MTS and 1 Ops manager! Come build the future of white-box audits with us: https://t.co/82Xp4vUak9

[Mario Giulianelli @glnmario]

Join our founding team to help us shape the future of frontier AI systems audits!

[Parallax @parallx_ai]

We're hiring founding technical staff and an operations manager in Europe/UK.

Join us in making white-box auditing a practical part of frontier AI evaluation: https://t.co/mr0lxojdek 10/

[Parallax @parallx_ai]

Introducing Parallax 🔍

We're a new UK/EU-based nonprofit research lab building scalable methods and infrastructure to audit the beliefs, goals, and plans that shape model behaviours in long-horizon agentic evaluations.

https://t.co/ge9pyOphtY — Thread 🧵 1/

[Parallax @parallx_ai]

We should not build increasingly capable AIs without comparable efforts to develop a rigorous understanding of their inner workings.

If these questions motivate you, we'd love to work together. 9/

https://t.co/oMCJJQQZY6

[Jacob Coxon @hilbertspaess]

If you are a lab researcher, I urge you to consider what the next few years will actually feel like. Do you want to kick off a superintelligent RL run without a rigorous understanding of its mind? Should you put your head down because “it’s happening anyway” - or take this moment to call for different conditions?