Papers5

#data efficiency

Causal World Modeling for Robot Control

Robots used to copy actions from videos without truly understanding how the world changes, so they often messed up long, multi-step jobs.

#robot world model#autoregressive diffusion#causal masking

On the Role of Discreteness in Diffusion LLMs

Intermediate

Ziqi Jin, Bin Wang et al.Dec 27arXiv

The paper asks what a truly good diffusion-based language model should look like and lists five must-have properties.

#diffusion language models#smooth corruption#discrete tokens

GenEnv: Difficulty-Aligned Co-Evolution Between LLM Agents and Environment Simulators

Intermediate

Jiacheng Guo, Ling Yang et al.Dec 22arXiv

GenEnv is a training system where a student AI and a teacher simulator grow together by exchanging tasks and feedback.

#GenEnv#co-evolutionary learning#difficulty-aligned curriculum

OpenDataArena: A Fair and Open Arena for Benchmarking Post-Training Dataset Value

Intermediate

Mengzhang Cai, Xin Gao et al.Dec 16arXiv

OpenDataArena (ODA) is a fair, open platform that measures how valuable different post‑training datasets are for large language models by holding everything else constant.

#OpenDataArena#post-training datasets#data-centric AI

VOYAGER: A Training Free Approach for Generating Diverse Datasets using LLMs

Intermediate

Avinash Amballa, Yashas Malur Saidutta et al.Dec 12arXiv

VOYAGER is a training-free way to make large language models (LLMs) create data that is truly different, not just slightly reworded.

#VOYAGER#determinantal point process#dataset diversity