Lumos Agents (AI2)

non-profit

https://allenai.github.io/lumos/

Activity Feed Request to join this org

AI & ML interests

None defined yet.

Recent Activity

yuchenlin authored a paper about 2 months ago

On Memorization of Large Language Models in Logical Reasoning

DaYin authored a paper 4 months ago

CogVideoX: Text-to-Video Diffusion Models with An Expert Transformer

yuchenlin authored a paper 6 months ago

WildGuard: Open One-Stop Moderation Tools for Safety Risks, Jailbreaks, and Refusals of LLMs

View all activity

ai2lumos's activity

yuchenlin

authored a paper about 2 months ago

On Memorization of Large Language Models in Logical Reasoning

Paper • 2410.23123 • Published Oct 30 • 18

DaYin

authored a paper 4 months ago

CogVideoX: Text-to-Video Diffusion Models with An Expert Transformer

Paper • 2408.06072 • Published Aug 12 • 37

yuchenlin

authored 2 papers 6 months ago

WildGuard: Open One-Stop Moderation Tools for Safety Risks, Jailbreaks, and Refusals of LLMs

Paper • 2406.18495 • Published Jun 26 • 12

MantisScore: Building Automatic Metrics to Simulate Fine-grained Human Feedback for Video Generation

Paper • 2406.15252 • Published Jun 21 • 14

DaYin

authored a paper 6 months ago

ChatGLM: A Family of Large Language Models from GLM-130B to GLM-4 All Tools

Paper • 2406.12793 • Published Jun 18 • 31

yuchenlin

authored a paper 6 months ago

WildVision: Evaluating Vision-Language Models in the Wild with Human Preferences

Paper • 2406.11069 • Published Jun 16 • 13

yuchenlin

authored 2 papers 7 months ago

Magpie: Alignment Data Synthesis from Scratch by Prompting Aligned LLMs with Nothing

Paper • 2406.08464 • Published Jun 12 • 65

WildBench: Benchmarking LLMs with Challenging Tasks from Real Users in the Wild

Paper • 2406.04770 • Published Jun 7 • 27

yuchenlin

posted an update 7 months ago

Post

2386

Introducing **BaseChat**!
https://huggingface.co/spaces/allenai/BaseChat_URIAL

This is a demo for our URIAL paper that enables base LLMs to chat with in-context alignment. You can talk directly with base, untuned LLMs to find out what knowledge and skills they have already learned from pre-training instead of SFT or xPO or RLHF. Also, you can use this to explore the pre-training data of base LLMs by chatting! I found a very interesting case: Base version of Llama-3-8B often thinks it is built by OpenAI, lol.