Sonia Murthy

Excited to be presenting our work on using cognitive models to interpret pluralistic values in LLMs once again as a spotlight talk 🌟 at the NeurIPS CogInterp workshop! Come by upper level room 5AB today and check out the paper here: arxiv.org/abs/2506.20666

@SaleemaAmershi @adamfourney @ASwearngin77874 @bansalg_ @HsseinMzannar @HuaWenyue31539 @w_epperson @ZacharyHuang12 @MayaMurad0 @ecekamar @HosnRafa hi all! I'll be at neurips and would love to learn more about the phd internships on your team, especially any projects in human-AI interaction and safety. my dms should be open 🙂

Reach out to me or these amazing humans to learn more about our team or #MSR AI Frontiers! @adamfourney @ASwearngin77874 @bansalg_ @HsseinMzannar Tyler Payne @HuaWenyue31539 @w_epperson @ZacharyHuang12 @MayaMurad0 @ecekamar @HosnRafa

📢We're hiring! Join an incredible team building AI agents that work *with* people and contribute meaningfully to society. Details below 👇 P.S. I'll be at #NeurIPS2025 and #WiML this week. DM me to chat about agents🤖 or #MSR AI Frontiers!

bruce is great at making research resources and this one has been a huge help for my human studies in the stream! check it out ✨

My rockstar MATS mentee @BruceWLee2 has just open-sourced his sleek and elegant codebase for AI control research, ppl should give it a try!

@sarahcat21 Hi Sarah! I just gave a talk today where I proposed versions of each of these directions, so was really surprised to see this pop up on my feed - I’ll be at NeurIPS and would love to chat!

I'll be among dozens (hundreds?) of VCs attending NeurIPS this year, but among the few who might be more interested in topics like managing episodic memory with RL, avoiding model collapse when training with synthetic data, and more effectively using base models to guide exploration, than who is leading your seed round at $1B post. So ping me if you want to chat :)

@aurielws @NeurIPSConf I’d love to join!

Doing a small Applied AI research dinner at @NeurIPSConf in San Diego with me and some other friends at other big labs (Gemini, Anthropic, Open AI) and Applied AI companies.  Can folks help recommend a friend (or yourself) that I should be inviting this year? I love meeting new people at #Neurips and the food will be comped 🙂

📝 New paper! Two strategies have emerged for controlling LLM behavior at inference time: in-context learning (ICL; i.e. prompting) and activation steering. We propose that both can be understood as altering model beliefs, formally in the sense of Bayesian belief updating. 1/9

Zach did a stellar job on our new paper looking at what recipes make for language models that are representationally aligned with humans! Read his tweetprint and recruit him for grad school!

@ZachStuddiford @siddsuresh97 @kushin_m hi this is cool work! I might be biased because I worked on something that has a very similar spirit arxiv.org/abs/2506.20666, but was excited to see y'all support the motivations and importance we saw around this kind of LLM analysis 😀

Thanks to @siddsuresh97, @kushin_m, and Tim Rogers for their support on this work! I’m excited to be applying to grad school this cycle so reach out if you want to chat about this Code 💻: github.com/Knowledge-and-… Paper 📄: arxiv.org/abs/2510.01030 10/10

We’re drowning in language models — there are over 2 mil. of them on Huggingface! Can we use some of them to understand which computational ingredients — architecture, scale, post-training, etc. – help us build models that align with human representations? Read on to find out 🧵

@kiran_tomlinson hey Kiran! couldn’t message you but I’d love to learn more about these openings/projects if you have some time to chat this week? 🙂

My team at Microsoft Research is hiring PhD interns for next summer! If you’re interested in understanding or improving human-LLM systems, apply here: jobs.careers.microsoft.com/global/en/shar… Topics we’re studying include LLM personalization, reasoning, collaboration, benchmarking, ++

Thanks to my lovely collaborators @rosieyzh, @_jennhu, @ShamKakade6, @m_wulfmeier, Peng Qian, and @TomerUllman and the Kempner Institute! 🧠 [end]

We also trace the evolution of value trade-offs during alignment by evaluating model checkpoints for 8 unique base model x feedback dataset x alignment algorithm. We see the largest shifts in values early on in training, with strongest effects of base model choice.

Excited to present our new paper as a spotlight talk 🌟 at the Pragmatic Reasoning in LMs workshop at #COLM2025 this Friday! 🍁 Come by room 520B @ 11:30am tomorrow to learn more about how LLMs' pluralistic values evolve over reasoning budgets and alignment 🧵

In our new paper, we ask whether language models solve compositional tasks using compositional mechanisms. 🧵

Presenting this today (5/1) at the 4pm poster session (Hall 3) at #NAACL2025! Come chat about alignment, personalization, and all things cognitive science 🐟

NEW blog post: Do modern #LLMs capture the conceptual diversity of human populations? #KempnerInstitute researchers find #alignment reduces conceptual diversity of language models. Read more: bit.ly/4hNjtiI @soniakmurthy @tomerullman @_jennhu

Many thanks to my collaborators and @KempnerInst for helping make this idea come to life!🌱

(9/9) Code and data for our experiments can be found at: github.com/skmur/onefish-… Preprint: arxiv.org/abs/2411.04427 Also, check out our feature in the @KempnerInst Deeper Learning Blog! bit.ly/417WVDL

(1/9) Excited to share my recent work on "Alignment reduces LM's conceptual diversity" with @TomerUllman and @jennhu, to appear at #NAACL2025! 🐟 We want models that match our values...but could this hurt their diversity of thought? Preprint: arxiv.org/abs/2411.04427