hopeschroeder.bsky.social
Studying NLP, CSS, and Human-AI interaction. PhD student @MIT. Previously at Microsoft FATE + CSS, Oxford Internet Institute, Stanford Symbolic Systems
hopeschroeder.com
23 posts
229 followers
228 following
Regular Contributor
Conversation Starter
comment in response to
post
Fascinating and important work!!
comment in response to
post
100%. I see this reasoning quickly proliferating across the social sciences as LLMs are integrated into text analysis tasks. Concerning- and would love to discuss sometime!
comment in response to
post
Read our 📎Preprint for more, incl practical guiding questions: arxiv.org/abs/2501.12557
Congrats to @rockpang.bsky.social on great leadership, and grateful to have learned so much from co-authors Kynnedy Smith, @s010n.bsky.social, @ziangxiao.bsky.social, @emtseng.bsky.social, and Danielle Bragg
comment in response to
post
We find LLM-related contributions in all domains such as health, productivity, writing, and design. As LLMs continue to permeate software systems, we need to update the way we discuss and evaluate them.
comment in response to
post
What role did the LLM play in your project? How did you disclose the models and prompts? What are the potential limitations of using LLMs for your selected role? (For example: How will the performance of the LLM-powered research tool affect the validity of your research?)
comment in response to
post
Researchers should think about how the LLM’s role in the work relates to potential limitations of the system AND resulting claims made in a study of that system, particularly when using or studying closed models. Here are a few q's to consider:
comment in response to
post
85% of these contributions use or study closed, proprietary models, amplifying challenges to reproducibility, transparency regarding behavior and performance, and further evaluation.
comment in response to
post
In this review of 153 LLM-related papers at CHI from 2020-24, we identified 5 roles that LLMs play: LLMs as system engines (62%), Users’ Perceptions of LLMs (24%), LLMs as objects of study (10%), LLMs as research tools (9%), & LLMs as participants & users (7%).
comment in response to
post
local in-person gatherings with friends, particularly focused on casual creativity and crafting, help me stay grounded and optimistic during these difficult times. How about you?
comment in response to
post
5) Which travel tricks or products have I missed? Always looking to make this easier!
comment in response to
post
4) I travel with blank thank you notes! There are always people to thank along the way :)
comment in response to
post
3) Charging cords are chaotic. This magnetic battery pack for my iPhone works wonders. A friend gave me this cute one: velvetcaviar.com/products/ber...
comment in response to
post
2) In addition to an N-95 face mask, (nothing like getting sick to ruin a trip) I've also started bringing hydrating sheet masks for my face during to deal with dehydration- helps me feel fresh and sane!
comment in response to
post
This looks very cool! I’ll share with my lab, which works on storytelling through voice. (Ccc.mit.edu)
comment in response to
post
Assistance in generating a pesky table is different than generating a whole abstract... right? Curious if anyone has used this and what you think!
comment in response to
post
Tom Costello ( @tomcostello.bsky.social ) has made freely available tools that let you build LLM interaction into Qualtrics
publish.obsidian.md/qualtrics-do...
comment in response to
post
Second: Generative Agent Simulations of 1,000 People (arxiv.org/abs/2411.10109) uses long qualitative interviews to simulate personas with very high correlation with traditional survey results.