frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

Show HN: I built a web tool to see and edit what an AI thinks before it answers

https://lucid.earthpilot.ai
12•ada1981•4h ago
I run a small AI lab and playground and got super excited about Anthropics paper "Verbalizable Representations Form a Global Workspace in Language Models" (https://transformer-circuits.pub/2026/workspace/index.html)

It talks about how they use a tool they call a Jacobian Lens to view inside the middle layers of LLM while it's working before it commits to a word (token).

I wanted to see if I could get a version of this running on the open models and to my surprise it worked! I ran some experiments with it and build a public facing free tool anyone can use with your own prompts.

Ask the model to describe a symbol of "three curving lines of water" and you can watch "ocean", "sea", and "surf" light up a few layers deeper before it settles on "waves".

You can also edit the internal state. Insert "fire" into the middle layer of the ocean prompt and the answer shifts to something about heat.

For fun / curiosity sake, I also developed way to let the model read its own inner workspace and then decide to suppress or amplify a concept, and run the prompt again.

Interesting finding from running it across models. J-lens beats a plain logit lens on some architectures and does nothing on others, and it isn't about size. A 0.5B Qwen reads better than a 2.8B Pythia. Every Pythia I tried gained basically nothing; the Llama and Qwen models gained a lot. https://lucid.earthpilot.ai/research

This is a 48 hour old project based on emerging research and built on a small model, a small probe set on rented GPUs - but I found it genuinely exciting. The code is open.

I also included a page context "Docent" AI agent you can chat with about whatever you see to help understand what is going on.

Happy to have folks poke around and break it.

I imagine the applications for allowing models to self-reflect / edit internal states can be useful for alignment, confidence, bias detection, etc. and this tool lets you play with the early stages of that.

Comments

ada1981•22m ago
also from our lab is Personality Bench: https://persona.earthpilot.ai/ which runs every frontier model through personality tests.
krackers•5m ago
>J-lens beats a plain logit lens on some architectures and does nothing on others, and it isn't about size

The paper talked about this, the jacobian matrix corrects for the shift in basis from initial to final layer compared to logit lens which assumes that the residual remains in the same basis across layers. Maybe the latter is true for models/architectures so the J-lens doesn't do anything extra?

GPT-5.6

https://openai.com/index/gpt-5-6/
875•logickkk1•5h ago•650 comments

Show HN: Getting GLM 5.2 running on my slow computer

https://github.com/JustVugg/colibri
144•vforno•14h ago•27 comments

EU Parliament greenlights Chat Control 1.0

https://www.patrick-breyer.de/en/eu-parliament-greenlights-chat-control-1-0-breyer-our-children-l...
858•rapnie•11h ago•410 comments

Show HN: 18 Words

https://18words.com/
743•pompomsheep•9h ago•269 comments

Hy3

https://hy.tencent.com/research/hy3
319•andai•6h ago•72 comments

Train sim created by just one person is being called the best ever made

https://kotaku.com/a-train-sim-created-by-just-one-person-is-being-called-the-best-ever-made-2000...
134•oumua_don17•4d ago•58 comments

No leap second will be introduced at the end of December 2026

https://datacenter.iers.org/data/latestVersion/bulletinC.txt
204•ChrisArchitect•8h ago•161 comments

A road to Lisp: Why Lisp

https://scotto.me/blog/2026-07-09-why-lisp/
77•silcoon•9h ago•75 comments

I Changed My Name

https://robida.net/entries/2026/07/01/i-changed-my-name
22•surprisetalk•3d ago•20 comments

Postgres rewritten in Rust, now passing 100% of the Postgres regression tests

https://github.com/malisper/pgrust
222•SweetSoftPillow•15h ago•275 comments

Girls just wanna have fast MPMC queues with bounded waiting

https://nahla.dev/blog/waitfree_queue/
101•EvgeniyZh•3d ago•20 comments

The glass backbone: Why the Army's logistics will break in the next war

https://mwi.westpoint.edu/the-glass-backbone-why-the-armys-logistics-will-break-in-the-next-war/
241•baud147258•8h ago•323 comments

A possible future for Damn Interesting

https://www.damninteresting.com/a-possible-future/
195•mzur•6h ago•21 comments

Launch HN: Context.dev (YC S26) – API to get structured data from any website

https://www.context.dev
61•TheYahiaBakour•6h ago•47 comments

GLM 5.2 is nearly as accurate as a human book keeper

https://toot-books.pages.dev/blog/glm-5-2-vat-benchmark
152•adamkurkiewicz•3h ago•92 comments

Muse Spark 1.1

https://ai.meta.com/blog/introducing-muse-spark-meta-model-api/
289•ot•8h ago•161 comments

Wildcard (YC W25) Is Hiring a Founding Engineer

https://www.ycombinator.com/companies/wildcard/jobs/ZSLVaaU-founding-engineer
1•kaushikmahorker•5h ago

How to Start a Ruby Meetup

https://guides.rubyevents.org/meetups/
43•mooreds•3h ago•13 comments

Meta reuses old RAM in new servers with custom bridge chip

https://www.theregister.com/systems/2026/06/29/zuck-saves-meta-bucks-by-reusing-memory-from-old-s...
276•ihsw•6d ago•201 comments

TLS certificates for internal services done right

https://tuxnet.dev/posts/tls-for-internal-services/
110•mrl5•7h ago•79 comments

Show HN: Rubiks Cube Solver

https://speedcube.com.br/
7•wozzp•1h ago•0 comments

Buried Apple feature turns an iPhone into the perfect kids' dumb phone

https://www.wired.com/story/this-buried-apple-feature-turns-an-iphone-into-the-perfect-kids-dumb-...
216•PotatoNinja•3d ago•140 comments

How GitHub gave every repository a durable owner

https://github.blog/security/application-security/how-github-gave-every-repository-a-durable-owner/
12•ascertain•1h ago•4 comments

Interview with Mitchell Hashimoto about Ghostty and Zig

https://alexalejandre.com/programming/interview-with-mitchell-hashimoto/
5•veqq•4h ago•0 comments

SimPolitics: America’s quest to solve politics with computers

https://mitpress.mit.edu/9780262053198/simpolitics/
55•mckelveyf•7h ago•8 comments

Almost Always Unsigned

https://graphitemaster.github.io/aau/
37•gavide•1d ago•48 comments

Opinionated and easy Pi.dev configuration

https://lazypi.org/
87•lwhsiao•6h ago•57 comments

Show HN: I mapped 8.5M research papers into an interactive atlas

https://tomesphere.com/atlas
51•leonickson•19h ago•17 comments

Show HN: I built a web tool to see and edit what an AI thinks before it answers

https://lucid.earthpilot.ai
12•ada1981•4h ago•2 comments

ChatGPT Work

https://openai.com/index/chatgpt-for-your-most-ambitious-work/
297•Tiberium•5h ago•143 comments