[2503.08980] I Predict Therefore I Am: Is Next Token Prediction Enough to Learn Human-Interpretable Concepts from Data?

Computer Science > Machine Learning arXiv:2503.08980 (cs) [Submitted on 12 Mar 2025 (v1), last revised 28 Feb 2026 (this version, v4)] Title:I Predict Therefore I Am: Is Next Token Prediction Enough to Learn Human-Interpretable Concepts from Data? Authors:Yuhang Liu, Dong Gong, Yichao Cai, Erdun Gao, Zhen Zhang, Biwei Huang, Mingming Gong, Anton van den Hengel, Javen Qinfeng Shi View a PDF of the paper titled I Predict Therefore I Am: Is Next Token Prediction Enough to Learn Human-Interpretable Concepts from Data?, by Yuhang Liu and 8 other authors View PDF HTML (experimental) Abstract:Recent empirical evidence shows that LLM representations encode human-interpretable concepts. Nevertheless, the mechanisms by which these representations emerge remain largely unexplored. To shed further light on this, we introduce a novel generative model that generates tokens on the basis of such concepts formulated as latent discrete variables. Under mild conditions, even when the mapping from the latent space to the observed space is non-invertible, we establish rigorous identifiability result: the representations learned by LLMs through next-token prediction can be approximately modeled as the logarithm of the posterior probabilities of these latent discrete concepts given input context, up to an linear transformation. This theoretical finding: 1) provides evidence that LLMs capture essential underlying generative factors, 2) offers a unified and principled perspective for understanding...