How AI Really Thinks: New Research Reveals LLMs Build Meaning in Real Time
Researchers discovered that AI language models don't generate text by retrieving pre-formed answers—they actively construct meaning as they write, with each word reshaping what comes next. The finding could reshape how companies evaluate AI reliability, design better safety guardrails, and understand why LLMs sometimes produce inconsistent or unpredictable outputs.
Originaltitel: Proto-Interpretation: The Temporality of Large Language Model Inference
We show that autoregressive generation in large language models exhibits a temporal structure: each token is not only conditioned on the past but also reshapes the future continuation space. We call this process proto-interpretation : the probabilistic redistribution across competing continuations through which the model gradually commits to one emerging branch of meaning. Using a minimal ambiguity case, we demonstrate branch competition and sequential commitment during inference. These findings reveal meaning in LLMs as a dynamic, temporally unfolding process, shifting interpretability from static model states to inference-time dynamics.