… This paper introduces InternLM2, an open-source LLM that … The pre-training process of
InternLM2 is meticulously detailed, … In this paper, we introduce InternLM2, a new Large …

… In this paper, we introduce InternLM2.5-StepProver which improves its automated
theorem-proving ability via large-scale expert iteration and achieves state-of-the-art on multiple …

Creating AI systems that can interact with environments over long periods, similar to human
cognition, has been a longstanding research goal. Recent advancements in multimodal …

… Our model based on InternLM27B [77] not only significantly outperforms existing multimodal
models but also matches or even surpasses GPT-4V [58] and Gemini Pro [76] in certain …

… We employ InternLM2-Chat as our foundation model and perform a two-stage supervised
fine-tuning (SFT) to specialize it for legal domain. The training pipeline is presented in Figure 2…

… A comprehensive analysis of InternLM2.5-StepProver is conducted on several standard
formal benchmarks, in comparison with our previous model InternLM2-StepProver, as well as a …

… This project is based on the base model InternLM2.57B-chat and has developed a
lightweight psychological counseling large language model application with low hardware …

… This paper introduced Xmodel-2, a 1.2-billion-parameter model optimized for reasoning
tasks. By leveraging the maximal update parametrization (µP), Warmup-Stable-Decay (WSD) …

… In this subsection, we assess the performance of InternLM2-Chat-20B-200K, which utilizes
three distinct retrievers on Law Reading scenarios. Results illustrated in Table 5, indicate a …

… 16 out of 28 benchmarks based on InternLM2-7B [143] backend. As shown in Figure 1, the
performance of IXC2.5 matches or even surpasses proprietary APIs, eg, GPT4V [112] and …