A naukri.com initiative
Medium
3w
313
Image Credit: Medium
A dive into how pass@k is calculated for evaluation of LLM’s coding
- pass@k is a metric used for evaluating the coding capability of LLM's.
- It generates k code solutions per problem.
- An alternative approach to calculate an unbiased estimator is proposed.
- Enlarging the generated samples helps reduce variance and improve accuracy.
Read Full Article
18 Likes
For uninterrupted reading, download the app