Search papers, labs, and topics across Lattice.
This paper investigates the distinction between detecting the utility of an auxiliary signal and effectively learning to act on it in the context of LLM acquisition agents. The authors establish a reward-SNR floor that dictates when it is feasible to learn per-instance acquisition policies, demonstrating that learned routing strategies fail to outperform random selection across various granularities. Using Structured Hypothesis Embeddings (SHE), the study shows that while the method can generate ranked intent hypotheses, its effectiveness is conditional on the model architecture and dataset regime, revealing that learned acquisition strategies collapse below the established SNR threshold.
Detecting a signal's average benefit doesn't guarantee that agents can learn to act on it, with a critical reward-SNR floor determining success.
Many pipelines can pay a per-example cost to acquire an auxiliary, model-derived observation -- an LLM's structured reasoning, a slow oracle, an expensive measurement -- and then must decide when the acquired signal is worth using. Our thesis is a distinction that is easy to miss: detecting that such a signal helps on average is not the same as learning to act on it per instance, and a reward-SNR floor governs when the second is even possible. Even when the signal is faithful and an in-sample oracle picking the top-b examples by realized reward shows a sizable apparent gain, no deployable policy can learn when to acquire it: across per-impression, cluster, regime, and uplift-tree granularities, learned routing never beats random, and a matched-moment noise placebo reproduces>=100% of the oracle's apparent gain -- the apparent"learnable structure"is order statistics of noise. We explain this with one distinction, detecting a mean effect vs. learning a per-instance acquisition policy, and a reward-SNR detectability floor: routing is estimable offline only if the reward SNR rho clears rho*(N) ~= 2.8/sqrt(N), with a positive control confirming a true low-SNR limit rather than a broken pipeline. As a concrete instantiation we introduce Structured Hypothesis Embeddings (SHE): a frozen LLM turns a user history into ranked, confidence-scored, evidence-grounded intent hypotheses, fused into a recommender. On three public datasets (MIND, REES46, Amazon-Beauty), SHE is faithful and calibratable, yet its value is backbone- and regime-conditional (significant over an ordered GRU, +0.0114, 95% CI [+0.0030, +0.0209], but a global redundancy gap indistinguishable from zero), and learned acquisition collapses at every granularity because all three datasets sit below the floor. The realizable unit is a design-time regime gate, not a per-instance policy. We release code and a one-command reproduction.
