ProfHumphreys
@profhumphreys.bsky.social
Reposted by ProfHumphreys
New preprint from @lauraknelson.bsky.social, @mattwilkens.bsky.social, and myself tests different ways of simulating the past with LLMs. We don't fully answer the title question here—just show that simple strategies based on prompting and fine-tuning are insufficient. +
Can Language Models Represent the Past without Anachronism?
Before researchers can use language models to simulate the past, they need to understand the risk of anachronism. We find that prompting a contemporary model with examples of period prose does not pro...
arxiv.org
May 2, 2025 at 12:47 PM
New preprint from @lauraknelson.bsky.social, @mattwilkens.bsky.social, and myself tests different ways of simulating the past with LLMs. We don't fully answer the title question here—just show that simple strategies based on prompting and fine-tuning are insufficient. +
Reposted by ProfHumphreys
There are notable gender differences in the way people view AI, but these gaps are more pronounced among experts we surveyed: 63% of men in this group say AI’s impact on the U.S. over the next two decades will be very or somewhat positive, compared with 36% of women.
April 9, 2025 at 4:02 PM
There are notable gender differences in the way people view AI, but these gaps are more pronounced among experts we surveyed: 63% of men in this group say AI’s impact on the U.S. over the next two decades will be very or somewhat positive, compared with 36% of women.
Reposted by ProfHumphreys
✨New paper ✨
Introducing 🌍MultiBLiMP 1.0: A Massively Multilingual Benchmark of Minimal Pairs for Subject-Verb Agreement, covering 101 languages!
We present over 125,000 minimal pairs and evaluate 17 LLMs, finding that support is still lacking for many languages.
🧵⬇️
Introducing 🌍MultiBLiMP 1.0: A Massively Multilingual Benchmark of Minimal Pairs for Subject-Verb Agreement, covering 101 languages!
We present over 125,000 minimal pairs and evaluate 17 LLMs, finding that support is still lacking for many languages.
🧵⬇️
April 7, 2025 at 2:56 PM
✨New paper ✨
Introducing 🌍MultiBLiMP 1.0: A Massively Multilingual Benchmark of Minimal Pairs for Subject-Verb Agreement, covering 101 languages!
We present over 125,000 minimal pairs and evaluate 17 LLMs, finding that support is still lacking for many languages.
🧵⬇️
Introducing 🌍MultiBLiMP 1.0: A Massively Multilingual Benchmark of Minimal Pairs for Subject-Verb Agreement, covering 101 languages!
We present over 125,000 minimal pairs and evaluate 17 LLMs, finding that support is still lacking for many languages.
🧵⬇️
Reposted by ProfHumphreys
*Excellent* paper. There's so much potential here for building simulations as a major social science research method. It's only going to grow over the next decade. Now is the time to build it on solid ground with deep roots.
Should we use LLMs 🤖 to simulate human research subjects 🧑? In our new preprint, we argue sims can augment human studies to scale up social science as AI technology accelerates. We identify 5 tractable challenges and argue this is a promising and underused research method 🧪🧵 arxiv.org/abs/2504.02234
LLM Social Simulations Are a Promising Research Method
Accurate and verifiable large language model (LLM) simulations of human research subjects promise an accessible data source for understanding human behavior and training new AI systems. However, resul...
arxiv.org
April 5, 2025 at 5:50 PM
*Excellent* paper. There's so much potential here for building simulations as a major social science research method. It's only going to grow over the next decade. Now is the time to build it on solid ground with deep roots.
Reposted by ProfHumphreys
The idea that you shouldn't be expected to know what data you're using to train your systems and that doing so is "an impossible task" is so normalized that its hard to know that this was not always the case even in the field of AI.
March 18, 2025 at 4:43 PM
The idea that you shouldn't be expected to know what data you're using to train your systems and that doing so is "an impossible task" is so normalized that its hard to know that this was not always the case even in the field of AI.
Reposted by ProfHumphreys
Statement on academic freedom, diversity, equity and inclusion at the Journal of Consumer Research:
consumerresearcher.com/academic-fre...
consumerresearcher.com/academic-fre...
Academic Freedom, Diversity, Equity and Inclusion at the Journal of Consumer Research - Journal of Consumer Research
The Journal of Consumer Research (JCR) is an independent academic journal that remains dedicated to improving academic freedom, diversity, equity,
consumerresearcher.com
March 15, 2025 at 10:16 PM
Statement on academic freedom, diversity, equity and inclusion at the Journal of Consumer Research:
consumerresearcher.com/academic-fre...
consumerresearcher.com/academic-fre...
Is anyone else also finding the yacht rock of Michael McDonald soothing these days? open.spotify.com/track/5GvWrv...
I Keep Forgettin' (Every Time You're Near)
Michael McDonald · If That's What It Takes · Song · 1982
open.spotify.com
February 17, 2025 at 5:20 PM
Is anyone else also finding the yacht rock of Michael McDonald soothing these days? open.spotify.com/track/5GvWrv...
Is anyone on Pixelfed? Because I am there as well and want friends. Fine me at AshleeHumphreys
January 26, 2025 at 2:39 AM
Is anyone on Pixelfed? Because I am there as well and want friends. Fine me at AshleeHumphreys
Does ‘made with love’ sell? Research reveals who values handmade products the most
theconversation.com/does-made-wi...
theconversation.com/does-made-wi...
Does ‘made with love’ sell? Research reveals who values handmade products the most
‘Mindful’ consumers feel handmade goods contain more ‘love’ and are willing to pay more for them. For faster-paced ‘doers’, handmade marketing can actually backfire.
theconversation.com
January 23, 2025 at 8:54 PM
Does ‘made with love’ sell? Research reveals who values handmade products the most
theconversation.com/does-made-wi...
theconversation.com/does-made-wi...
Reposted by ProfHumphreys
New paper: Do social media algorithms shape affective polarization?
We ran a field experiment on X/Twitter (N=1,256) using LLMs to rerank content in real-time, adjusting exposure to polarizing posts. Result: Algorithmic ranking impacts feelings toward the political outgroup! 🧵⬇️
We ran a field experiment on X/Twitter (N=1,256) using LLMs to rerank content in real-time, adjusting exposure to polarizing posts. Result: Algorithmic ranking impacts feelings toward the political outgroup! 🧵⬇️
November 25, 2024 at 8:32 PM
New paper: Do social media algorithms shape affective polarization?
We ran a field experiment on X/Twitter (N=1,256) using LLMs to rerank content in real-time, adjusting exposure to polarizing posts. Result: Algorithmic ranking impacts feelings toward the political outgroup! 🧵⬇️
We ran a field experiment on X/Twitter (N=1,256) using LLMs to rerank content in real-time, adjusting exposure to polarizing posts. Result: Algorithmic ranking impacts feelings toward the political outgroup! 🧵⬇️
Hello to my first follower! @consumerlife.bsky.social ❤️❤️❤️
November 19, 2024 at 6:06 PM
Hello to my first follower! @consumerlife.bsky.social ❤️❤️❤️