发现大模型共用计数子电路，跨任务精准倒数
A Shared Subcircuit Lets LLMs Count Down Across Tasks
- 识别出大模型中通用的倒数子电路，实现跨任务计数
- 在固定长度句末任务中，模型准确预测剩余词数
- 适用于隐含目标长度的任务，适合研究模型泛化机制
写一个恰好十二个词的句子；在正确的密码子处结束一段DNA序列；格式化一个ASCII表格。这些任务都需要模型跟踪距离目标还有多少词。本文在 Llama-3.1-70B-Instruct 中发现一种通用机制：一个‘倒数子电路’，能将当前位置与目标长度对比，估算剩余时间。我们首先在控制环境下隔离该子电路，任务是写出固定长度的句子并以指定词结尾。接着分析其表征几何结构，发现该子电路使用了此前在前沿大模型上识别出的相同特征模式，表明该模式具有跨模型共享性。最后，通过无监督探测，在自然语言数据集中发现多种使用该子电路的任务，包括目标长度需从上下文推断的情况。研究表明，逆向工程子电路有助于理解行为如何从单个例子泛化到多个任务甚至不同模型。
原文摘要 · Abstract (English)
Writing a sentence of exactly twelve words; ending a DNA sequence at the right codon; formatting an ASCII table. These are all tasks that language models can do that requires tracking how many tokens remain before a target. In this work, we identify in Llama-3.1-70B-Instruct a general mechanism for performing these tasks: a "countdown subcircuit" that compares the current position to a goal length and estimates the time remaining until then. We first isolate a countdown subcircuit in a controlled setting, in which the model is tasked with writing a fixed-length sentence ending in a specified word. We then investigate the geometry of the representations used by the subcircuit, and find that the subcircuit uses an identical motif previously identified in a frontier LLM on a separate task, thus suggesting that this motif is shared across models. Finally, we use unsupervised probing on a natural language dataset to find a variety of other tasks where this subcircuit is used, including tasks where the goal length is inferred from context rather than explicitly stated. Our work suggests that reverse-engineering subcircuits allows us to understand how behaviors generalize from a single example to many different tasks and even models.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。