Method
Self-consistency / best-of-N
Sampling several answers and choosing by voting or reranking.
7 cases from 7 companies apply Self-consistency / best-of-N. 3 report results in production.
Typical tasks: Evaluate models, Label data with LLMs, Rank search results, Retrieve and rank from millions of items and 1 more. Companies: HSE University, Netflix, Etsy, Yandex, Spotify, SberDevices and 1 more. Often combined with: LLM-as-a-judge, Supervised fine-tuning, Uncertainty quantification, Golden set / benchmark, Chain-of-thought and 22 more.
Doubling Medical Safety: Fine-Tuning Open LLMs for Women's Health Without Human Labels![]() | ![]() | healthtech | 2026-06 |
| When Models Should Stay Silent | general | 2026-05 | |
| Evaluating Netflix Show Synopses with LLM-as-a-Judge | media-streaming | 2026-04 | |
| How Etsy Uses LLMs to Improve Search Relevance | marketplace | 2026-01 | |
Estimating Generative Model Confidence in Data Labeling![]() | ![]() | general | 2025-05 |
| Building Confidence: A Case Study in How to Create Confidence Scores for GenAI Applications | fintech | 2024-12 | |
RLHF for retrieval based NLP system![]() | ![]() | general | 2023-05 |


