← 研踪 PaperTrailarXiv:2611.04515 · v1
Learnings and Pitfalls from AI-Assisted A/B Experimentation at Scale
首次提交 2026-11-03 16:41
版本更新 2026-11-03 16:41
雷达观测 2026-11-04 23:39
摘要
A/B testing underpins product decisions by providing trustworthy, reproducible quantification of proposed changes' effects. Growing large language model (LLM) capabilities have motivated efforts to automate experimentation workflows. Yet LLMs are nondeterministic and prone to hallucination, properties that conflict with the trust experimentation systems must provide. Experiments are also operationally messy: diagnosing anomalies, debugging configurations, and conducting bespoke follow-up analyses require flexible reasoning and manual effort. These tasks let agents accelerate experimentation without replacing its deterministic statistical foundations. We report lessons from building and operating such an agent. Our central finding is that this tension is resolved not by making the model deterministic, but by bounding where its nondeterminism can reach; the model selects and sequences procedures, while deterministic code computes every quantity it reports. Three consequences surprised us: withholding information from the model improved its answers, demand concentrated on mundane retrieval rather than the autonomous decisions we anticipated, and the agent's eagerness to help threatened statistical validity.
作者备注：5 pages, 1 figure, 2 tables. Accepted to the Conference on Digital Experimentation at MIT (CODE@MIT 2026)
观测到的版本
相邻记录可能跳过中间版本；首次发现 v2 及以上也会标为版本更新。
v1 · 2026-11-03 16:41 · 最新观测版本
Learnings and Pitfalls from AI-Assisted A/B Experimentation at Scale
A/B testing underpins product decisions by providing trustworthy, reproducible quantification of proposed changes' effects. Growing large language model (LLM) capabilities have motivated efforts to automate experimentation workflows. Yet LLMs are nondeterministic and prone to hallucination, properties that conflict with the trust experimentation systems must provide. Experiments are also operationally messy: diagnosing anomalies, debugging configurations, and conducting bespoke follow-up analyses require flexible reasoning and manual effort. These tasks let agents accelerate experimentation without replacing its deterministic statistical foundations. We report lessons from building and operating such an agent. Our central finding is that this tension is resolved not by making the model deterministic, but by bounding where its nondeterminism can reach; the model selects and sequences procedures, while deterministic code computes every quantity it reports. Three consequences surprised us: withholding information from the model improved its answers, demand concentrated on mundane retrieval rather than the autonomous decisions we anticipated, and the agent's eagerness to help threatened statistical validity.
查看此版本解读 →查看此版本原文 ↗