Beyond the Parameter Monolith: Reconstructive Memories, Executable Skills, and Residual Assembly for Language Models
We investigate FEM-ASM, a finite-element-method-inspired organization in which independently constructed document states and deterministic executable skills contribute typed proposals to a shared language-model state.
ProofPaper ↗
Key points
- Language-model systems can separate contextual computation, persistent storage, and exact execution instead of updating all capabilities through one shared parameter system.
- An explicit residual operator reconciles proposals attached to common interface nodes.
- An attention-free Multi-Mesh prototype learns causal language modeling but does not establish competitive general capability.
- Support-aware lexical indices make these elements addressable under provenance-controlled query construction.
Sources (1)
- [1]Beyond the Parameter Monolith: Reconstructive Memories, Executable Skills, and Residual Assembly for Language ModelsHugging Face Daily Papers · Oct 2, 12:00 AM
We investigate FEM-ASM, a finite-element-method-inspired organization in which independently constructed document states and deterministic executable skills contribute typed proposals to a shared language-model state.
Language-model systems can separate contextual computation, persistent storage, and exact execution instead of updating all capabilities through one shared parameter system.
Extractive summary: sentences quoted from the sources.