Valid Stopping for LLM Generation via Empirical Dynamic Formal Lift
This work addresses the efficiency and reliability of stopping LLM generation for users in NLP applications, though it is incremental as it builds on existing sequential testing methods.
The paper tackled the problem of when to stop language model generation by introducing Sequential-EDFL, which uses anytime-valid sequential testing to control error rates, resulting in a 22-28% reduction in generation length compared to baselines while maintaining formal guarantees with 12% computational overhead.
We introduce Sequential-EDFL (Empirical Dynamic Formal Lift), applying anytime-valid sequential testing to language model generation stopping. Our approach tracks information lift -- the log-likelihood ratio between full models and deliberately weakened "skeleton" baselines -- using self-normalized empirical-Bernstein e-processes that provide formal delta-level error control regardless of stopping time. We handle unknown centering through online mean estimation, combine multiple parameters via mixture e-processes, and support adaptive resets under distributional drift. On six benchmarks, Sequential-EDFL reduces generation by 22-28% vs. sequential baselines while maintaining delta-level control with 12% computational overhead. We introduce automated skeletons (distilled submodels, randomized logits) and show robustness across skeleton families. Composing EDFL with a lightweight correctness gate (sentence boundaries + verifier) improves end-task correctness while preserving anytime-valid guarantees by only delaying stopping. Our certificates control information sufficiency, not factual correctness -- 10.9% of stopped sequences remain incorrect even with the gate (13.2-22.7% without it). EDFL serves as a first-stage filter reducing verification burden by 83%, not as a standalone solution for safety-critical domains.