Daniel Hershcovich
@daniel_hers
Followers
990
Following
2K
Media
55
Statuses
372
Assistant Professor @coastalcph @DIKU_Institut, University of Copenhagen. Interests: meaning, multiculturalism and promoting responsible behavior with #NLProc
Copenhagen, Denmark
Joined October 2008
New preprint: "Decoding Alignment: A Critical Survey of LLM Development Initiatives through Value-setting and Data-centric Lens" 🔍 Beyond a survey, this is a presentation of a series of concerns on what we could frame as "corporate" alignment #llms #ai
arxiv.org
AI Alignment, primarily in the form of Reinforcement Learning from Human Feedback (RLHF), has been a cornerstone of the post-training phase in developing Large Language Models (LLMs). It has also...
0
2
11
New Preprint 🎉: "Cultural Evaluations of Vision-Language Models Have a Lot to Learn from Cultural Theory". We review recent works on culture in VLMs and argue for deeper grounding in cultural theory to enable more inclusive evaluations. Paper 🔗: https://t.co/9AoRHTFG58
1
10
88
This talk and panel are starting now - crucial anthropological insights for NLP!
Next, we have Lise Dobrin, Professor at the Department of Anthropology and Linguistics Program at the University of Virginia. She will share insights gained through nearly thirty years work with Arapesh people in the Pacific island nation of Papua New Guinea.
0
0
5
This incredible workshop is starting now! Join at room "Pecos" if you're at @naaclmeeting
📣 Excited about our Cross Cultural Considerations in NLP workshop! @c3_nlp @ @naaclmeeting coming up in no time—just two weeks to go! Full program here: https://t.co/YZK3EcaU1W Co-org. w/ @sunipa17 @LucianaBenotti @daniel_hers @vinodkpg @YongCaoPlus @LiZhou21 @IfeAdebara ❤️
0
0
0
@YongCaoPlus Haijiang Liu, @rnav_arora @IAugenstein @paul_rottger @daniel_hers Specializing Large Language Models to Simulate Survey Response Distributions for Global Populations. https://t.co/QiWSbOOYLS
#NAACL2025 #NLProc
https://t.co/RR6wvw6x4m
arxiv.org
Large-scale surveys are essential tools for informing social science research and policy, but running surveys is costly and time-intensive. If we could accurately simulate group-level survey...
🎉 Excited to share our NAACL main paper on simulating survey response distributions via first-token alignment fine-tuning. Many thanks to our collaborators: Haijiang Liu, @rnav_arora @IAugenstein @paul_rottger and @daniel_hers. More details, paper link and code, see below👇
1
2
6
📣 Excited about our Cross Cultural Considerations in NLP workshop! @c3_nlp @ @naaclmeeting coming up in no time—just two weeks to go! Full program here: https://t.co/YZK3EcaU1W Co-org. w/ @sunipa17 @LucianaBenotti @daniel_hers @vinodkpg @YongCaoPlus @LiZhou21 @IfeAdebara ❤️
c3nlp.github.io
1
5
9
We are hiring a law+NLP postdoc in Copenhagen! Deadline April 7: https://t.co/YEB4PO7G59
0
2
11
🎉 Excited to share our NAACL main paper on simulating survey response distributions via first-token alignment fine-tuning. Many thanks to our collaborators: Haijiang Liu, @rnav_arora @IAugenstein @paul_rottger and @daniel_hers. More details, paper link and code, see below👇
5
3
16
📢Happy to share our paper "Does Mapo Tofu Contain Coffee? Probing LLMs for Food-related Cultural Knowledge" has been accepted to #NAACL2025 main! Work with Taelin Karidi, Wanlong Liu, @nGarneau @YongCaoPlus , Wenyu Chen, Haizhou Li, @daniel_hers
https://t.co/LKEc5FcEco
8
7
45
📣 C3NLP 2025 CFP now out 🎉 We’re thrilled to announce C3NLP will be at NAACL 2025 in Albuquerque, New Mexico! 🌎 📅 Workshop Details: https://t.co/HeOqMwXyx8 Co-organized w/ @vinodkpg @sunipa17 @LucianaBenotti @daniel_hers @YongCaoPlus @LauraCabelloPi @LiZhou21 @IfeAdebara
c3nlp.github.io
3
9
16
Prior work has used LLMs to simulate survey responses, yet their ability to match the distribution of views remains uncertain. Our new paper [ https://t.co/DleesiPbif] introduces a benchmark to evaluate how distributionally aligned LLMs are with human opinions. 🧵
4
38
160
Only 4 more days until the SICon 2024 submission deadline (August 29th)! Remember that we also accept abstracts about ongoing work as well as papers through ARR.
0
3
3
Had a great time co-organizing @c3_nlp this year w/ @YongCaoPlus @LiZhou21 @sunipa17 @LauraCabelloPi @daniel_hers @LucianaBenotti @IfeAdebara Packed room all day! Overheard someone saying c3nlp is trending! Didn't expect it back when we proposed it in 2022!
Amazing brochure to #ACL2024 🇹🇭 with an amazing 2nd edition of the workshop @c3_nlp 🗺️ with amazing co-organizers ♥️
2
6
42
Please provide us feedback about the workshop through our feedback form:
docs.google.com
Thank you for attending the 2nd workshop on Cross Cultural Considerations in NLP at ACL 2024. As we continue to develop this program, we would appreciate your input on how to make the workshop better!
0
1
3
Congratulations to the best paper authors! Thanks to everyone who attended and contributed. You can find our proceedings on
0
1
4
Join our industry panel discussion now and ask @moontae_lee and @vinodkpg about cross-cultural considerations at LG and Google! #ACL2024NLP
0
2
1
CreoleVal: Multilingual Multitask Benchmarks for Creoles Highlight: Evaluation of how accuracy changes under surface-level translation v.s. true localization! Paper Link: https://t.co/XWnN00h3Nt (Also features @stanfordnlp alum @ruthstrong_)
1
2
5
Thanks for a huge range of lightning talks on multilingual, multicultural, fun and diverse work! These will be presented as posters after lunch and our two amazing panels, so make sure to stick around! #ACL2024NLP
2
1
4
Many thanks Shalom, @aliceoh and @luischiruzzo for the excellent and mind-opening discussion, ending in a great conclusion - "don't despair"!
0
1
3