Alex Lyman

CS PhD student @ BYU
NLP guy
Free speech advocate
Retweets & Likes do not imply endorsement

[Alex Lyman @AlexMLyman]

If big companies have our data, can they use in conjunction with LLMs to persuade people on hot-button political issues? Our new study, published in PNAS, puts LLMs to the test persuading people on hot-button political topics. 🧵[1/6]

[Alex Lyman @AlexMLyman]

Bottom line: AI can shift opinions at scale, but personalization and chat didn’t add much. If you’re curious about the ability of AI to persuade-and its limits-check out our full open-access paper:
https://t.co/zClco9ozWz [6/6]

[Alex Lyman @AlexMLyman]

We tested four AI-driven persuasion strategies:
1️⃣ Generic message
2️⃣ Microtargeted message
3️⃣ Interactive direct arguments
4️⃣ Motivational interviewing
All were aimed at shifting people away from their starting views, but which worked best? 🤔 [2/6]

[Alex Lyman @AlexMLyman]

While AI is able to move people's attitudes towards the center, this does not increase "democratic reciprocity", a willingness to engage with people who hold opposing views. This suggests that merely changing policy opinions isn't enough.[5/6]

[Manoel @manoelribeiro]

Spaced repetition is a lie. Why don't I remember `tar -xvzf` after 10+ years?

[Pessimists Archive @PessimistsArc]

No one is ready for this. (1908)

[Alex Lyman @AlexMLyman]

Synthetic data is something else. (Taken from Cosmopedia v2)

[Karine Mellata @karine_exe]

We just caught our first "BANANA!"

[Alex Lyman @AlexMLyman]

I just had work on Language Models for quantifying eyewitness confidence statements in Psychological Science.

You can read the news brief here:
https://t.co/9OyghPDtoi

Or the full paper on SSRN:
https://t.co/mJiHHA8lsO

[Crémieux @cremieuxrecueil]

When I first saw this graph, the main thing that stood out to me was that Japan has the world's least spicy food.

Even Brits eat spicier food.

[Eric Hallahan @EricHallahan]

>Code fails to compile under CUDA Toolkit 12.1
>Upgrades CUDA Toolkit to 12.2
>Code fails to compile earlier than before

[Manoel @manoelribeiro]

One of our key sources of human data is no longer fully “human"!

We estimate that 33-46% of crowd workers on MTurk used large language models (LLMs) in a text production task - which may increase as ChatGPT and the like become more popular and powerful.

https://t.co/SJfKjDM6gX

[AI Asia Pacific Institute @AIAsiaPacific]

@pmddomingos argues that #AI regulation is misguided, and more on this podcast: https://t.co/zF3sVc5lmO

#artificialintelligence #generativeai #regulation #aiact #europe #podcast

[Frank A. Krueger @praeclarum]

“AI Safety” seams to be a euphemism for censorship and I don’t think big corps should be bragging about it quite so much.

[Tal Linzen @tallinzen]

I see someone on this website has come up with a specific number of training FLOPS above which a language model becomes AGI and needs to be regulated. Very glad I logged in this morning.

[Peyman Milanfar @docmilanfar]

advanced math in deep learning papers

[Alex Lyman @AlexMLyman]

Had a great time shoring up my reinforcement learning with @huggingface's course. Thanks @ThomasSimonini!