OLMoTrace: Tracing Language Model Outputs Back to Trillions of Training Tokens
This system helps users understand language model behavior for applications like fact-checking and hallucination detection, though it is incremental as it builds on existing infini-gram technology.
The researchers tackled the problem of tracing language model outputs to their training data by developing OLMoTrace, a system that identifies verbatim matches between model outputs and trillions of training tokens in real time, returning results within seconds.
We present OLMoTrace, the first system that traces the outputs of language models back to their full, multi-trillion-token training data in real time. OLMoTrace finds and shows verbatim matches between segments of language model output and documents in the training text corpora. Powered by an extended version of infini-gram (Liu et al., 2024), our system returns tracing results within a few seconds. OLMoTrace can help users understand the behavior of language models through the lens of their training data. We showcase how it can be used to explore fact checking, hallucination, and the creativity of language models. OLMoTrace is publicly available and fully open-source.