Thank you to everyone who joined the GeBNLP Workshop at #ACL2025! โจ
To help us improve future events, please share your feedback in our anonymous 5-minute survey. Your perspective is needed!
๐ Survey Link: forms.gle/mXtZ4mWUGSBN...
#GeBNLP #NLP
The Gender Bias in NLP workshop #ACL2025 is officially over, and we're buzzing with energy! โจ
A huge thank you to all our participants for a day packed with incredible discussions. You are the community we need ๐
See you next time!
#GeBNLP #NLP
It's time! โฐ Organizers are ready to kick off the Gender Bias in NLP Workshop #ACL2025. Join us for a day of talks, posters, and discussions!
๐๏ธ Today, August 1
๐ 9 AM - 6:15 PM
๐ Hall C
#GeBNLP #NLP
We are thrilled to announce our keynote speakers for the 6th Workshop on Gender Bias in NLP @ #ACL2025!
Please join us in welcoming:
๐นAnne Lauscher @a-lauscher.bsky.social
๐นMaarten Sap @maartensap.bsky.social
Full details: gebnlp-workshop.github.io/keynotes.html
See you on August 1! โ๏ธ
#NLP #GeBNLP
Our VLMs4All workshop is taking place today!
๐
on Thursday, June 12
โฒ๏ธ from 9AM CDT
๐๏ธin Room 104E
Join us today at @cvprconference.bsky.social for amazing speakers, posters, and a panel discussion on making VLMs more geo-diverse and culturally aware!
#CVPR2025
Do LLMs hallucinate randomly? Not quite.
Our #ACL2025 (Main) paper shows that hallucinations under irrelevant contexts follow a systematic failure mode โ revealing how LLMs generalize using abstract classes + context cues, albeit unreliably.
๐ Paper: arxiv.org/abs/2505.22630 1/n
๐๏ธ Save the date! It's official: The VLMs4All Workshop at #CVPR2025 will be held on June 12th!
Get ready for a full day of speakers, posters, and a panel discussion on making VLMs more geo-diverse and culturally aware ๐
Check out the schedule below!
๐จ Deadline Extension Alert for #VLMs4All Challenges! ๐จ
We have extended the challenge submission deadline
๐ ๏ธ New challenge deadline: Apr 22
Show your stuff in the CulturalVQA and GlobalRG challenges!
๐ sites.google.com/view/vlms4al...
Spread the word and keep those submissions coming! ๐โจ
Exciting release! AgentRewardBench offers that much-needed closer look at evaluating agent capabilities: automatic vs. human eval. Important findings here, especially on the popular LLM judges. Amazing work by @xhluca.bsky.social & team!
๐ Reminder & Call for #VLMs4All @ #CVPR2025!
Help shape the future of culturally aware & geo-diverse VLMs:
โ๏ธ Challenges: Deadline: Apr 15 ๐https://sites.google.com/view/vlms4all/challenges
๐ Papers (4pg): Deadline: Apr 22 ๐https://sites.google.com/view/vlms4all/call-for-papers
Join us!
Reviewers needed! ๐ข The 6th Workshop on Gender Bias in NLP at #ACL2025 (Vienna, Aug 1st) is looking for you! Sign up to review:
forms.gle/VkPU4vS4EacE... #NLP
Excited to be organizing the VLMs4All workshop at #CVPR2025! ๐
The workshop features fantastic speakers, a short-paper track, and two challenges, including one based on CulturalVQA. Donโt miss it!
6/ ๐ค Thanks to our steering committees and co-organizers for their hard work in making the VLMs4All Workshop possible!
@meharbhatia.bsky.social @rabiul.bsky.social @spandanagella.bsky.social @sivareddyg.bsky.social @svansteenkiste.bsky.social @karstanczak.bsky.social
Instruction-following retrievers can efficiently and accurately search for harmful and sensitive information on the internet! ๐๐ฃ
Retrievers need to be aligned too! ๐จ๐จ๐จ
Work done with the wonderful Nick and @sivareddyg.bsky.social
๐ mcgill-nlp.github.io/malicious-ir/
Thread: ๐งต๐
The potential for malicious misuse of LLM agents is a serious threat.
That's why we created SafeArena, a safety benchmark for web agents. See the thread and our paper for details: arxiv.org/abs/2503.04957 ๐
Llamas browsing the web look cute, but they are capable of causing a lot of harm!
Check out our new Web Agents โฉ Safety benchmark: SafeArena!
Paper: arxiv.org/abs/2503.04957
6/ @meharbhatia.bsky.social @hattie.bsky.social @nicolaspapernot.bsky.social @nicolaschapados.bsky.social @anamarasovic.bsky.social @ghadfield.bsky.social @sivareddyg.bsky.social
#AIAlignment #LLM #ResponsibleAI #NLP
5/ Read our full paper here: arxiv.org/abs/2503.00069
Letโs discuss! How should AI align with society? ๐ค๐ก
4/ We also discuss the role of participatory alignment, where diverse stakeholders help shape AI behavior rather than deferring solely to designers.
3/ Instead of perfecting rigid alignment objectives, we explore how LLMs can navigate uncertaintyโa feature, not a flaw!
2/ We propose leveraging societal alignment frameworks to guide LLM alignment:
๐น Social alignment: Modeling norms, values & cultural competence
๐น Economic alignment: Fair reward mechanisms & collective decision-making
๐น Contractual alignment: Legal principles for LLMs
1/ LLM alignment remains a challenge because human values are complex, dynamic, and often conflict with narrow optimization goals.
Existing methods like RLHF struggle with misspecified objectives.
๐ขNew Paper Alert!๐
Human alignment balances social expectations, economic incentives, and legal frameworks. What if LLM alignment worked the same way?๐ค
Our latest work explores how social, economic, and contractual alignment can address incomplete contracts in LLM alignment๐งต
The 6th edition of the Gender Bias in Natural Language Processing (GeBNLP) workshop is happening at #ACL2025 in Vienna!
๐ข Check out our Call for Papers! Find all the details on our website: gebnlp-workshop.github.io
We look forward to your submissions!
Great news! We've extended the direct submission deadline to *April 18*, giving you more time to prepare your submissions ๐
Check the new deadlines on our webpage: gebnlp-workshop.github.io/cfp.html
Presenting โจ ๐๐๐๐๐: ๐๐๐ง๐๐ซ๐๐ญ๐ข๐ง๐ ๐๐ก๐๐ฅ๐ฅ๐๐ง๐ ๐ข๐ง๐ ๐ฌ๐ฒ๐ง๐ญ๐ก๐๐ญ๐ข๐ ๐๐๐ญ๐ ๐๐จ๐ซ ๐๐ฏ๐๐ฅ๐ฎ๐๐ญ๐ข๐จ๐ง โจ
Work w/ fantastic advisors Dima Bahdanau and @sivareddyg.bsky.social
Thread ๐งต:
Screenshot of 'SHADES: Towards a Multilingual Assessment of Stereotypes in Large Language Models.' SHADES is in multiple grey colors (shades).
โซโช It's coming...SHADES. โชโซ
The first ever resource of multilingual, multicultural, and multigeographical stereotypes, built to support nuanced LLM evaluation and bias mitigation. We have been working on this around the world for almost **4 years** and I am thrilled to share it with you all soon.
Excited to present today a poster at @OECD in Paris @IASEAIorg based on our upcoming paper "Societal Alignment Frameworks Can Improve LLM Alignment" (stay tuned for the pre-print soon!๐). Today (Fri) at 1pm CET. Conference livestream: iaseai.org/conference
It turns out we had even more papers at EMNLP!
Let's complete the list with three more๐งต
๐โโ๏ธ๐
