telechat-a-comprehensive-large-language-model-with-extensive-pretraining-and-supervised-fine-tuning-for-conversational-ai

chatglm-rlhf-praktiken-zur-ausrichtung-großer-sprachmodelle-auf-menschliche-präferenzen


An efficient training-free attention control paradigm to mitigate entity leakage and attribute misalignment issues in text-to-image generation.



Enhancing Text-to-Image Generation Alignment through Attention Modulation
