🚀 Introducing 𝐖𝐨𝐥𝐟 🐺: a mixture-of-experts video captioning framework that outperforms GPT-4V and Gemini-Pro-1.5 in general scenes 🖼️, autonomous driving 🚗, and robotics videos 🤖. 👑: wolfv0.github.io/leaderboard.ht…

[CL] Training Large Language Models to Reason in a Continuous Latent Space S Hao, S Sukhbaatar, D Su, X Li… [Meta] (2024) arxiv.org/abs/2412.06769

COLM 2025: Oct 7-9 (conference), Oct 10 (workshops) Montreal, Canada Abstract deadline: March 20, 2025 Paper deadline: March 27, 2025 CFP: colmweb.org/cfp.html Workshop proposal deadline: April 14, 2025 Call: colmweb.org/cfw.html Mailing list: groups.google.com/g/colm-announc…

synthetic data is the weirdest name for data augmentation.

[LG] Task Vectors in In-Context Learning: Emergence, Formation, and Benefit L Yang, Z Lin, K Lee, D Papailiopoulos... [University of Wisconsin-Madison] (2025) arxiv.org/abs/2501.09240

New 3h31m video on YouTube: "Deep Dive into LLMs like ChatGPT" This is a general audience deep dive into the Large Language Model (LLM) AI technology that powers ChatGPT and related products. It is covers the full training stack of how the models are developed, along with mental

🚀Excited to share our #CHI2025 paper on AACessTalk, an #AI-driven tablet app that fosters communication between a minimally-verbal autistic (MVA) child and parent, which was an internship project of @NAVER_AI_LAB led by Dasom Choi! 🧵(1/6) #hci #nlproc #autism ACM CHI Conference

🧵(1/7)🚀 Excited to share our #CHI2025 preprint on ELMI, an LLM-infused interactive & intelligent sign language translation of lyrics for song-signing! Work led by Suhyeon Yoo as a research intern at NAVER AI LAB. #AI #Accessibility #HCI #NLProc

(🧵1/7) 🚀Excited to share our #CHI2025 preprint on ExploreSelf, an #LLM-guided tool supporting user-driven reflection and exploration of personal challenges, an internship project of NAVER AI LAB led by Inhwa! #HCI #AI #NLProc #MentalWellness #DigitalWellbeing

Our #CHI2025 paper on an AI-driven, contextual conversation mediation system for minimally verbal autistic children and parents received a 🏆Best Paper Award! This is the second CHI Best Paper of my group NAVER AI LAB after two years. So proud of Dasom Choi and coauthors.

Curious about why Knowledge Distillation works so well in generative models? In our latest paper, we offer a minimal working explanation! Please check it out!

Image-text alignment is hard — especially as multimodal data gets more detailed. Most methods rely on human labels or proprietary feedback (e.g., GPT-4V). We introduce: 1. CycleReward: a new alignment metric focused on detailed captions, trained without human supervision. 2.

Sharing interesting work by JangHyun . Estimating the importance of KV cache entries for pruning is challenging. Instead, we propose a simple approach: recover the context itself—akin to self-supervised learning (like MAE or BERT)—to obtain a generalized notion of importance.

Leaky Thoughts Hey AI devs, be careful how you prompt reasoning models. This work shows that reasoning traces frequently contain sensitive user data. More of my notes below:

🔎 Does Conversational SEO (C-SEO) actually work? Our new benchmark has an answer. Excited to announce C-SEO Bench: Does Conversational SEO Work? 🌐 RTAI: researchtrend.ai/papers/2506.11… 📄 Paper: arxiv.org/abs/2506.11097 💻 Code: github.com/parameterlab/c… 📊 Data: huggingface.co/datasets/param…