이진욱님의 블로그

https://arxiv.org/pdf/2411.04282Here's the text with each sentence on a new line: Large language models (LLMs) have shown impressive capabilities, but still struggle with complex reasoning tasks requiring multiple steps. While prompt-based methods like Chain-of-Thought (CoT) can improve LLM reasoning at inference time, optimizing reasoning capabilities during training remains challenging. We int..

https://arxiv.org/pdf/2411.06722v1

https://arxiv.org/pdf/2411.09595

https://arxiv.org/pdf/2411.07133

https://arxiv.org/pdf/2411.04997

https://arxiv.org/pdf/2410.02080

https://arxiv.org/pdf/2411.07232

https://arxiv.org/pdf/2411.08147Large language models (LLMs) have achieved substantial progress in processing long contexts but still struggle with long-context reasoning. Existing approaches typically involve fine-tuning LLMs with synthetic data, which depends on annotations from human experts or advanced models like GPT-4, thus restricting further advancements. To address this issue, we invest..

https://arxiv.org/abs/2405.17430

https://arxiv.org/pdf/2407.02392