Semantically Anchored Prompts for Relational Deep Learning

Recent work injects GNN embeddings of relational databases into frozen large language models (LLMs) as soft prompts through a learned projection. We show that such prompts are semantically blind: the projection places them off the manifold of pre-trained token embeddings, so the LLM cannot tell a customer from a transaction and its world knowledge is never activated. We propose RelAnchor, which learns an assignment instead of a projection. Schema Anchoring embeds the database's table and column descriptions with the frozen input embedding layer, yielding an on-manifold anchor basis that spans exactly the semantics of the database. Adaptive Composition expresses each row as a row-specific convex mixture over these anchors, gated with its graph representation, so that every prompt preserves relational distinctions while staying anchored to named schema concepts near the manifold. We prove that anchoring costs at most $O(1/\sqrt{k})$ in prompt fidelity with $k$ active anchors, governed by the LLM's Lipschitz constant on the compact on-manifold anchor hull rather than the unbounded ambient space. On RelBench, RelAnchor outperforms GNN-only baselines and the state-of-the-art graph-prompted LLM on the aggregate classification and regression metrics. Code is available at https://github.com/nmksjx/RelAnchor.