LongIns: A Challenging Long-context Instruction-based Exam for LLMs
This work addresses the need for better benchmarks to assess LLMs' actual long-context reasoning abilities, though it is incremental as it builds on existing instruction datasets.
The authors tackled the problem of evaluating long-context capabilities in large language models (LLMs) by proposing the LongIns benchmark, which revealed that top-performing models like GPT-4 with 128k context windows perform poorly on 16k evaluations and that many LLMs struggle with multi-hop reasoning under short contexts (less than 4k).
The long-context capabilities of large language models (LLMs) have been a hot topic in recent years. To evaluate the performance of LLMs in different scenarios, various assessment benchmarks have emerged. However, as most of these benchmarks focus on identifying key information to answer questions, which mainly requires the retrieval ability of LLMs, these benchmarks can partially represent the reasoning performance of LLMs from large amounts of information. Meanwhile, although LLMs often claim to have context windows of 32k, 128k, 200k, or even longer, these benchmarks fail to reveal the actual supported length of these LLMs. To address these issues, we propose the LongIns benchmark dataset, a challenging long-context instruction-based exam for LLMs, which is built based on the existing instruction datasets. Specifically, in our LongIns, we introduce three evaluation settings: Global Instruction & Single Task (GIST), Local Instruction & Single Task (LIST), and Local Instruction & Multiple Tasks (LIMT). Based on LongIns, we perform comprehensive evaluations on existing LLMs and have the following important findings: (1). The top-performing GPT-4 with 128k context length performs poorly on the evaluation context window of 16k in our LongIns. (2). For the multi-hop reasoning ability of many existing LLMs, significant efforts are still needed under short context windows (less than 4k).