Everything Everywhere All at Once: LLMs can In-Context Learn Multiple Tasks in Superposition
원문 초록
Large Language Models (LLMs) have demonstrated remarkable in-context learning (ICL) capabilities. In this study, we explore a surprising phenomenon related to ICL: LLMs can perform multiple, computationally distinct ICL tasks simultaneously, during a single inference call, a capability we term "task superposition". We provide empirical evidence of this phenomenon across various LLM families and scales and show that this phenomenon emerges even if we train the model to in-context learn one task at a time. We offer theoretical explanations that this capability is well within the expressive power of transformers. We also explore how LLMs internally compose task vectors during superposition. Furthermore, we show that larger models can solve more ICL tasks in parallel, and better calibrate their output distribution. Our findings offer insights into the latent capabilities of LLMs, further substantiate the perspective of "LLMs as superposition of simulators", and raise questions about the mechanisms enabling simultaneous task execution.
함께 살펴볼 논문
공통 태그를 근거로 연결했습니다. 실제 인용 관계나 논문의 우수성을 뜻하지 않습니다.
Assay2Mol: large language model-based drug design using BioAssay context
Is Diversity All You Need for Scalable Robotic Manipulation?
EAMamba: Efficient All-Around Vision State Space Model for Image Restoration
FineWeb2: One Pipeline to Scale Them All -- Adapting Pre-Training Data Processing to Every Language
Universal pre-training by iterated random computation
These are Not All the Features You are Looking For: A Fundamental Bottleneck In Supervised Pretraining
토론
이 논문과 연결된 포럼 글이 아직 없습니다.