Diffusion language models (DLMs) have attracted wide attention for their fast, parallel decoding, but they are believed to be far less efficient to train than autoregressive language models (ARMs): a widely cited study reports that a DLM needs roughly 16x the training compute of an ARM to match its likelihood, a figure...
Guang-Han Wang, B. He, Zhan Shi et al.· 3 citations
BadEvolve is proposed, a failure-aware task-poisoning framework against self-evolving LLM agents that diagnoses attack failures across four stages, i.e., Induction, Activation, Following, and Effect, and directs updates toward the failing stage.
Min-hua Lin, Zongyu Wu, Han-Qing Lu et al.· 0 citations
Large language models (LLMs) are increasingly used to automate AI research, including language-model (LM) development, offering a potential path toward recursive self-improvement. However, turning experimental feedback into a learned ability to propose effective improvements across models faces two challenges. First, t...
Min-hua Lin, Zi-Jun Wang, Ze-Wen Liu et al.· 2 citations
Language models as Research World Models (RWMs), which predict the outcomes of candidate interventions across research environments, are investigated andlations across 13 language models used as RWMs show that adding research knowledge can improve intervention ranking more than changing models or increasing reasoning e...
Zi-Jun Wang, Ze-Wen Liu, Min-hua Lin et al.· 3 citations
STAG is proposed, a stealthy trojan attack framework designed for the graph-language alignment interface of GFMs on TAGs that realizes trigger nodes as readable text through candidate retrieval and regularizes the trigger-attached subgraph so that its local structure remains close to the original subgraph.
Min-hua Lin, Zhi-Cheng Gao, Yilong Wang et al.· 0 citations
This is the first publicly reported autonomous post-training run at this scale, where prior public autonomous-ML-research demonstrations sit at GPT-2-class (~124M) budgets, and is treated as direct, auditable evidence that a scaled autonomous loop can produce discovery, not only optimization.
Zhan Shi, Bing-Xiang He, Yi-Si Sang et al.· arXiv.org· 3 citations
This formulation enables a systematic study of key self-improvement factors through the proposed Evo-Harness, and provides a principled understanding of how LLM agents can effectively learn on the fly.
Tian-Xin Wei, Zhan Shi, Min-hua Lin et al.· 14 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.