Sparsity Is Necessary: Polynomial-Time Stability for Agentic LLMs in Large Action Spaces

Abstract: Tool-augmented LLM systems expose a control regime that learning theory has largely ignored: sequential decision-making with a massive discrete action universe (tools, APIs, documents) in which only a small, unknown subset is relevant for any fixed task distribution. We formalize this setting as Sparse Agentic Control (SAC), where policies admit block-sparse representations over M >> 1 actions and rewards depend on sparse main effects and (optionally) sparse synergies. We study ell_{1,2}-regularized policy learning through a convex surrogate and establish sharp, compressed-sensing-style results: (i) estimation and value suboptimality scale as k (log M / T)^{1/2} under a Policy-RSC condition; (ii) exact tool-support recovery holds via primal-dual witness arguments when T > k log M under incoherence and beta-min; and (iii) any dense policy class requires Omega(M) samples, explaining the instability of prompt-only controllers. We further show that under partial observability, LLMs matter only through a belief/representation error epsilon_b, yielding an additive O(epsilon_b) degradation while preserving logarithmic dependence on M. Extensions cover tuning-free, online, robust, group-sparse, and interaction-aware SAC.