Shmulik Amar

🚨 RAG is a popular approach but what happens when the retrieved sources provide conflicting information?🤔 We're excited to introduce our paper: “DRAGged into CONFLICTS: Detecting and Addressing Conflicting Sources in Search-Augmented LLMs”🚀 A thread 🧵👇

🧵 New paper at Findings #ACL2025 @aclmeeting! Not all documents are processed equally well. Some consistently yield poor results across many models. But why? And can we predict that in advance? Work with Steven Koniaev and Jackie Cheung @Mila_Quebec @McGill_NLP #NLProc (1/n)

Check out our new paper on highly localized attributions, both in the input and the output!

🚨 Introducing LAQuer, accepted to #ACL2025 (main conf)! LAQuer provides more granular attribution for LLM generations: users can just highlight any output fact (top), and get attribution for that input snippet (bottom). This reduces the amount of text the user has to read by 2…

Excited to present our system demonstration paper on EventFull — an Event-Event Relation annotation tool — at #NAACL25 Come see us Thursday, May 1, at Poster Session I (16:00–17:30) (Paper and tool links at the end of the thread👇)

RefVNLI Towards Scalable Evaluation of Subject-driven Text-to-image Generation

LLMs struggle with tables—but how robust are they really? 🔍 ToRR goes beyond accuracy, testing real-world robustness across formats & tasks. 📊 Different formats, same data—models show brittle behavior affecting rankings. Prompt configuration is a key dimension for evaluation!🚀

🎉 I'm happy to share that our paper, Make It Count, has been accepted to #CVPR2025! A huge thanks to my amazing collaborators - @YoadTewel, @SegevHilit , @hirscheran, @RoyiRassin, and @GalChechik! 🔗 Paper page: make-it-count-paper.github.io Excited to share our key findings!

🚀 Big news for #VisualRiddles! We’re excited to announce that Visual Riddles has been accepted to the Creative AI Track at NeurIPS 2024! 🎉 Come explore our Visual Riddles Gallery—a showcase of cognitive and visual challenges for multimodal AI. 🧵

Happy to share our work "Counterfactual Generation from Language Models" with @AnejSvete, @vesteinns, and Ryan Cotterell! We tackle generating true counterfactual strings from LMs after interventions and introduce a simple algorithm for it. (1/7) arxiv.org/pdf/2411.07180

How diverse are the outputs of text-to-image models and how can we measure that? In our new work, we propose a measure based on LLMs and Visual-QA (VQA), and show NONE of the 12 models we experiment with are diverse. 🧵 1/11

Attending 🌴#EMNLP2024 or interested in what people are working on these days? We organized it all for you with Knowledge Navigator! Explore all @emnlpmeeting accepted papers mapped by themes and subtopics—giving you a bird’s-eye view of the conference knowledge-navigators.github.io/emnlp_presenta…

"Is It Really Long Context if All You Need Is Retrieval? Towards Genuinely Difficult Long Context NLP" was accepted to EMNLP 2024! arxiv.org/abs/2407.00402 🧵👇 If you're attending EMNLP this year, go chat with @omerNLP and @lovodkin93 about the future of long context evals!