Jason Yosinski
@jasonyo
Followers
7K
Following
2K
Media
57
Statuses
2K
Running experiments @OpenAI + @ml_collective. Prev: @Windscape_AI, Uber AI Labs founding team, adviser @RecursionPharma, Cornell, Montreal, Caltech 🌻
San Francisco, CA
Joined February 2008
@ManasJoglekar @GabrielDWu1 @j_asminewang @boazbaraktcs @mia_glaese See also @techreview writeup https://t.co/M53lXKS3fk
OpenAI has trained its LLM to confess to bad behavior
0
2
6
This was a fun project (that I jumped into halfway through) with @ManasJoglekar, Jeremy Chen, @GabrielDWu1, @j_asminewang, @boazbaraktcs, and @mia_glaese. Boaz wrote a good casual summary: https://t.co/IqWSkhCYQD
1/5 Excited to announce our paper on confessions! We train models to honestly report whether they “hacked”, “cut corners”, “sandbagged” or otherwise deviated from the letter or spirit of their instructions. @ManasJoglekar Jeremy Chen @GabrielDWu1 @jasonyo @j_asminewang
1
3
10
We just posted a blog + paper on a a simple but effective approach to model honesty called "Confessions" TL; DR: normal RL training rewards for high performance on a task. Confession training is a separate phase that rewards only for honesty. Test look promising! More:
In a new proof-of-concept study, we’ve trained a GPT-5 Thinking variant to admit whether the model followed instructions. This “confessions” method surfaces hidden failures—guessing, shortcuts, rule-breaking—even when the final answer looks correct. https://t.co/4vgG9wS3SE
1
2
18
Today, OpenAI is launching a new Alignment Research blog: a space for publishing more of our work on alignment and safety more frequently, and for a technical audience. https://t.co/n3oIhyDZHd
39
137
1K
xAI could be the next Elon Musk company to go public. What separates it from the rest: • $230B pre-money valuation talks after merging with X, more than 4× in under a year, driven by institutional demand and strategic insiders. • 200,000 Nvidia GPUs deployed in ~120 days,
1
5
48
A little while ago, many of you gave generously to support a number of MLC-Nigeria researchers in attending Deep Learning Indaba #DLI2025. Here's the crew 👇that attended; from what we hear it was a bustle of talks, posters, mentorship, and sparks of collaboration!
Today we say goodbye to @DeepIndaba after six inspiring days in Kigali rich with keynotes, tutorials, workshops, mentorship circles, and insightful posters that kept us learning non-stop. Some of us were only able to make it down to #DLI2025 because of your generous support.
1
15
67
Today we say goodbye to @DeepIndaba after six inspiring days in Kigali rich with keynotes, tutorials, workshops, mentorship circles, and insightful posters that kept us learning non-stop. Some of us were only able to make it down to #DLI2025 because of your generous support.
The opportunity gap in AI is more striking than ever. We talk way too much about those receiving $100M or whatever for their jobs, but not enough those asking for <$1k to present their work. For 3rd year in a row, @ml_collective is raising funds to support @DeepIndaba attendees.
1
17
94
Help send a bunch of researchers to DL Indaba this year! For less than one H100 we can send 25 people!
The opportunity gap in AI is more striking than ever. We talk way too much about those receiving $100M or whatever for their jobs, but not enough those asking for <$1k to present their work. For 3rd year in a row, @ml_collective is raising funds to support @DeepIndaba attendees.
0
20
37
At AEG, we don’t just create events—we create memories. Join the team behind unforgettable music and sports moments
0
5
31
Starting in 1 hour: @thebasepoint presents Anthropic's "Biology of a Large Language Model" work at the DLCT reading group. Paper: https://t.co/gxg8Ixlud2 Come for the chain of thought, stay for the rabbits and habbits. Zoom info below 👇
3
4
13
Starting in 30 min!
Next Research Jam is in 14 hours, tomorrow morning at 8am PT. Stop by this virtual lab meeting to hear research ideas and updates on projects in progress! Zoom info at https://t.co/bmwCfHTkJR
0
1
4
Next MLC Research Jam is tomorrow; sharing two ideas myself to mix things up :)
Next Research Jam is in 14 hours, tomorrow morning at 8am PT. Stop by this virtual lab meeting to hear research ideas and updates on projects in progress! Zoom info at https://t.co/bmwCfHTkJR
1
1
6
Starting in 15 min!
This week at Deep Learning: Classics and Trends we're kicking off a new five part mini-series on LLM Interpretability. Up first: @thesubhashk shows how LLMs represent numbers on a helix and use it to add! Join Friday at 10am PT, zoom here: https://t.co/f5VxrVu4Mg
0
1
4
Our equine entertainers are the true stars of the show! 🐴
0
1
26
I am sitting here watching my HF smolagent slowly reason about and click on Captcha squares one a time 🙈. Is this general AI?
1
0
6
Tomorrow at 10am PT we'll have our next MLC OpenClubHouse, our 25th 🎉! Stop by to hang out, catch up with friends, and chat about ML or anything else. We'll meet in the MLC Discord #openclubhouse channel: https://t.co/gVKOZbNCVh
discord.com
Discord is great for playing games and chilling with friends, or even building a worldwide community. Customize your own space to talk, play, and hang out.
0
2
6
If you're at #ICLR2025, stop by the ML Collective Picnic Lunch on Monday at 12:30, graciously hosted by Alex Bezzubov! All welcome, bring your own lunch and meet new research friends and collaborators. 🥰 https://t.co/wDqKTB2r99
luma.com
ML Collective open picnic lunch at ICLR 2025 It's been a conference or few since our last in-person lunch, so let's get together and eat! Everyone welcome, no…
0
1
10
✨Our new @unireps paper tries to answer why the Lottery Ticket Hypothesis (LTH) fails to work for different random inits through the lens of weight-space symmetry. We improve the transferability of LTH masks to new random inits leveraging weight symmetries. 🧵(1/6)
7
26
85
And...there we go! TL;DR: we are launching a new event series called "Industry Round Tables" with its first instance on Thursday, August 22! Register here if interested: https://t.co/vPn7Oy6yAX
Get ready for our first LinkedIn post!
1
4
18
We speed up renewable energy site selection & due diligence, reducing months of work to minutes. With $11M in new funding led by @navitascapital, we'll improve our software tools to allow quick, informed decisions that accelerate the energy transition. https://t.co/7qGkIho9XX
2
6
13
Base assets are now usable in everyday life. Top up your Tria card, tap to pay globally, and keep full custody. Use creator coins anywhere Visa or Mastercard are accepted. Onchain meet real world.
414
243
746
I had a pretty fun conversation with @JonKrohnLearns the other day on startups, wind energy, the electrical grid in the US, ML, and (of course) how neural networks really work :)
One of my all-time favorite A.I. researchers, Dr. Jason Yosinski (@jasonyo), is my guest today! He details how his startup is using ML to collect wind energy more efficiently and digs into visualizing/understanding deep neural networks. Watch here: https://t.co/esAWRX44MZ
1
3
14
Canadians, if you‘re considering switching to a heat pump to heat and cool your home, and you’re curious about utility bill costs, pay back period, and reduction in greenhouse gas emissions, I made a thing for you. Link 👇
1
5
33