I've spent a two-week vacation to fine-tune a large language model on my writing from the last decade to produce what I lovingly call #IAN (intelligence artificielle neuronale). I wrote a Substack post about what it is and how I made it! Check it out :)
Follow
Jan Hendrik Kirchner
@janhkirchner
phd student in comp neuroscience @ mpi brain research frankfurt, universalprior.substack.com
Science & TechnologySan Francisco, CAJoined March 2018
Jan Hendrik Kirchner’s Tweets
We’re developing a new tool to help distinguish between AI-written and human-written text. We’re releasing an initial version to collect feedback and hope to share improved methods in the future.
635
2,656
8,616
Super excited to share an initial version of a tool to help distinguish AI-written and human-written text! Would love to hear feedback for improving the tool in the future!
3
9
72
Everyone has a right to know whether they are interacting with a human or AI.
Language models like ChatGPT are good at posing as humans.
So we trained a classifier to distinguish between AI-written and human-written text.
But it's not fully reliable.
24
28
155
Show this thread
Awesome work by , , and many others!
It's been amazing to see how quickly it came together!
2
1
8
Show this thread
(I know the link doesn't look legit, I promise it's legit! All the better domain names were taken.)
Show this thread
The post has a bunch of pretty pictures and schematics, but really is just a thinly veiled attempt to get people to label data for me. Check out fashionator.xyz and tell me about your aesthetic preferences!
1
Show this thread
I wrote about fashion and ended up with a deep(-ish) meditation on all things pretty! Check it out!
1
1
7
Show this thread
would the past have been less dangerous if the atomic bomb had been released incrementally? (slowly ramping up in units of equivalent tnt)
1
3
has written a beautiful piece about the procrastination support group that he organized for me
2
4
This one was a ton of fun to write and work on in general!
Quote Tweet
@janhkirchner has written a beautiful piece about the procrastination support group that he organized for me
universalprior.substack.com/p/simulator-mu
3
A subset A ⊆ ℝ² is called a *cloud* around 𝑥 if every line through 𝑥 has a finite intersection with A.
[Komjáth 2001]: The following are equivalent:
(1) Three clouds cover ℝ².
(2) The continuum hypothesis holds.
27
198
1,344
The application of ChatGPT I'm most happy about is that finally my parents have heard of where I work
37
109
4,748
We’ve trained language models to be better at responding to adversarial questions, without becoming obtuse and saying very little. We do this by conditioning them with a simple set of behavioral principles via a technique called Constitutional AI: anthropic.com/constitutional
22
301
1,204
Show this thread
ChatGPT is incredibly limited, but good enough at some things to create a misleading impression of greatness.
it's a mistake to be relying on it for anything important right now. it’s a preview of progress; we have lots of work to do on robustness and truthfulness.
919
4,216
28.7K
Show this thread
🙏 to my colleagues for making it feel like i get to walk into the modern bell labs every day.
please consider joining us at openai! being in the room for research breakthroughs is awesome, the problems are interesting, the people are great, and we ship.
youtu.be/AyOnug-3OKM
134
199
4,135
While I still haven't gotten back into the rhythm of writing regularly (🥲) my good friend Sam is much better at it and is bringing some very exciting research discussions to paper. Highly recommended! snellessen.substack.com/p/introduction
3
New post in our Best of Science Blogging feed - “Drug Addicts and Deceptively Aligned Agents - a Comparative Analysis” by the excellent and Nadia Montazeri
4
9
If everything's a big manifold, why do neurons often code for human-interpretable factors? In arxiv.org/abs/2210.01768 we show the most efficient biological representation puts different factors in different neurons. This lets us build machines that disentangle too! 1/9
6
157
652
Show this thread
New Paper! "What are the Red Flags for Neural Network Suffering?" by and . Big thanks to gardeners , , , , , and many others for their reviews!
1
9
18
Show this thread
At EAG SF and interested in AI safety? Stop by OpenAI at the career fair or DM me. We're hiring across teams (incl. Trust and Safety, Security) - and we are always interested in hearing what you want to work on & what you think we should work on!
1
7
24
it’s been a blast reading Ava’s thoughts on this! She has a fantastic knack at getting art out of DALLE
Quote Tweet
Over the past few weeks, I've been wondering about the impact & utility of DALL-E & related algorithms.
My awesome friend @janhkirchner just joined @OpenAI & I got to play around with DALL-E.
With that, here's a piece I wrote for
Jan's substack. Enjoy!
universalprior.substack.com/p/hello-dall-e
1
2
Over the past few weeks, I've been wondering about the impact & utility of DALL-E & related algorithms.
My awesome friend just joined & I got to play around with DALL-E.
With that, here's a piece I wrote for
Jan's substack. Enjoy!
2
4
"fully aligned intelligence, photorealistic 4k"
(DALL-E 2)
1
2
25
(I've quietly published a short post on that already some days ago - just another reason to subscribe to my Substack universalprior.substack.com/p/a-quick-one- )
2
Show this thread
When I was young(er) I started coding because I wanted to build AI. That's pretty difficult, so I pivoted to "being part of the team that builds AGI". Now I'm happy to announce that I'm approaching my goal - I've joined OpenAI (Alignment Team) 🥳 Looking forward to exciting times
8
99
Show this thread
(unless you're working in a field like physics or mathematics where all the low-hanging might *actually* have already been picked)
Show this thread
I strongly disagree with this argument - and I think a lot of people have a terrible time with their research because they believe that "easy"="bad". There is no a priori reason for why that should be true.
universalprior.substack.com/p/on-scaling-a
Quote Tweet
If you are never getting stuck and your research always moves smoothly from beginning to end, then you probably aren't challenging yourself enough. Easy problems are good warm-up for harder problems, but the goal isn't to solve lots of easy problems.
Show this thread
1
Show this thread
Young researchers, check out this event 👇
Quote Tweet
I'm beyond excited (basically ecstatic) to share that the IICCSSS summer school in Computational Cognitive Science is happening again this year in September! Registration is open now :) Check it out! iiccsss.org
1
2
I'm beyond excited (basically ecstatic) to share that the IICCSSS summer school in Computational Cognitive Science is happening again this year in September! Registration is open now :) Check it out! iiccsss.org
6
15
Really looking forward to working with the legendary Scott Aaronson!
11
47
355
I had a blast working with you on this <3
Quote Tweet
If you asked yourself "Shouldn't there be more people with a neuroscience/cognitive science background in AI Safety?" already, then my new post written together with the wonderful @janhkirchner might be of interest to you: snellessen.com/2022/06/15/bra! #aisafety #neuroscience
2
Our latest alignment research looks at ways AI can help humans evaluate language model output more effectively
31
154
420
This thing has been an absolute blast (esp. the figures)! The dataset is published out now. Thanks to everyone involved :)
Quote Tweet
New paper & AI alignment dataset 
We collected and cataloged AI alignment research literature and analyzed the resulting dataset in an unbiased way to identify major research directions. We hope the dataset can be used to build tools for AI alignment.
lesswrong.com/posts/FgjcHiWv
We collected and cataloged AI alignment research literature and analyzed the resulting dataset in an unbiased way to identify major research directions. We hope the dataset can be used to build tools for AI alignment.
lesswrong.com/posts/FgjcHiWvShow this thread
3
And I'm also using the occasion to dive a bit deeper into the neuroscience that motivated this project in this weeks essay :) Lots of pretty pictures in there, like this timelapse of the developing zebrafish sensory system! universalprior.substack.com/p/the-brain-th
GIF
Quote Tweet
Super happy to share this with the research world! During brain development, function (synapses) emerges simultaneously with structure (dendrites, axons etc). Here we take a stab at disentangling how those factors interact! twitter.com/GjorJulijana/s…
2
2
Super happy to share this with the research world! During brain development, function (synapses) emerges simultaneously with structure (dendrites, axons etc). Here we take a stab at disentangling how those factors interact!
Quote Tweet
Continuing our work on how synaptic inputs organize on dendrites, we added growth and generate dendritic morphologies with approx optimal wiring. Just in time for the #embo #dendrites conference in beautiful Crete. By the talented @janhkirchner and Lucas. biorxiv.org/content/10.110
2
5
Couldn't get that thread out of my head, so I processed the argument into maths! Optimal control and extreme value theory show how the difficulty of adversarial attacks only increases linearly while their (supposed) probability decreases exponentially universalprior.substack.com/p/adversarial-
Quote Tweet
! twitter.com/0xdoug/status/…Show this thread
1
Foucault: a lotta shit is in fact like prison
Rorty: all inquiry is more arbitrary than we like to think [can't fit less banal v.]
Fisher: Many people feel weird about capitalism, this deserves close attention
Bourdieu: dude I'm not continental I'm just difficult and French wtf
3
4