Agentic Control in Variational Language Models
DOI: 10.48550/arXiv.2604.12513
Abstract
We study whether a variational language model can support a minimal and measurable form of agentic control grounded in its own internal evidence. Our model combines local variational hidden computation (EVE), a homeostatic latent regulator, structurally aware checkpoint retention and a calibrated uncertainty-aware controller operating on top of the retained model. Rather than treating uncertainty as a passive diagnostic measured after prediction, we treat it as an operational signal that can regulate training, support checkpoint retention and guide inference-time intervention. The resulting framework is deliberately focused. It studies a closed-loop form of internal control in which structural and predictive signals become actionable. Empirically, the variational backbone improves over a matched deterministic reference on the language-modeling task while also exhibiting a richer and more usable uncertainty profile. On top of this backbone, the calibrated controller remains active, uses multiple actions under a full agentic evaluation and yields a positive quality-cost trade-off. These results support a precise claim: internal uncertainty can serve not only as a descriptive property of a variational language model, but also as a practical control interface for regulation, checkpoint retention and minimal agentic routing.
Keywords
- agentic control
- variational language models
- internal uncertainty
- EVE
- checkpoint retention
- uncertainty-aware routing
Citation
Ruffenach, Yves. “Agentic Control in Variational Language Models.” arXiv preprint, April 14, 2026. DOI: 10.48550/arXiv.2604.12513.
Related research
Variational neuron · Variational Transformer · Internal uncertainty