# Measure Learning at Steady State: A BIRD-SQL Formula 1 Case Study

arXiv: https://arxiv.org/abs/2609.31640 · HTML: https://mlsanity.com/paper/2609.31640/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 4 of its 6 references that are in the map.
- First posted on arXiv 11 Sep 2026.

## Builds on (most cited first, 4 of 4)

- [Can LLM Already Serve as A Database Interface? A BIg Bench for Large-Scale Database Grounded Text-to-SQLs](https://mlsanity.com/paper/2305.03111/) (2023)
- [Cartridges: Lightweight and general-purpose long context representations via self-study](https://mlsanity.com/paper/2506.06266/) (2025)
- [Continual Learning Bench: Evaluating Frontier AI Systems in Real-World Stateful Environments](https://mlsanity.com/paper/2606.05661/) (2026)
- [AgentCL: Toward Rigorous Evaluation of Continual Learning in Language Agents](https://mlsanity.com/paper/2606.02461/) (2026)

## Abstract (arXiv)

Continual Learning Bench scores learning as short-horizon gain versus a reset baseline and finds naive full-context ICL strongest among the memories it tested. We treat ICL as one learning system and score it on a longer shared-world schedule. Steady-state learning is the gap versus baseline on a pre-set late window (last 40 of 174 BIRD-SQL formula-1 questions). We split the score into exploration efficiency (SQL probes), task reward (hits), and delivery cost (API dollars and context size). On gpt-5.6-luna, late probes fall from 4.6-5.6 to 0.95 while hits rise only modestly and ICL context grows to about 95k tokens with cost roughly doubling. Short-horizon gain understates the late probe saving and misses the cost inversion, so we find that unbounded ICL is a poor candidate for the learning mechanism.