Linke Ouyang

TL;DR: Experimental results demonstrate the superiority of InternLM-XComposer2 based on InternLM2-7B in producing high-quality long-text multi-modal content and its exceptional vision-language understanding performance across various benchmarks, where it not only significantly outperforms existing multimodal models but also matches or even surpasses GPT-4V and Gemini Pro in certain assessments.

TL;DR: InternLM2 is an open-source LLM that outperforms its predecessors in comprehensive evaluations across various tasks, including long-context modeling and open-ended subjective evaluations. It utilizes innovative pre-training and optimization techniques to capture long-term dependencies and achieve remarkable performance on the ``Needle-in-a-Haystack" test.

TL;DR: This paper introduces a novel solution, Hallucination-Aware Direct Preference Optimization (HA-DPO), which reframes the hallucination problem as a preference selection task and proposes an efficient pipeline for constructing positive~(non-hallucinatory) and negative~(hallucinatory) sample pairs, ensuring a high-quality, style-consistent dataset for robust preference learning.