[docs] add evaluations math

[aolemila]

Resolves #265

[gemini-code-assist]

Summary of Changes

Hello @aolemila, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request introduces new documentation that outlines the process and results of evaluating the DeepSeek-R1-Distill-Qwen-1.5B model on the MATH-500 dataset, demonstrating the impact of various server configurations and prompt strategies. Simultaneously, it refines the core sampling mechanisms within the JAX runtime, specifically for TPU execution, by implementing a new sorting-based approach for top_k_top_p_min_p sampling and streamlining conditional logic for improved consistency.

Highlights

• New Evaluation Documentation: A new documentation file, docs/evaluations/evals.md, has been added. This file provides comprehensive details on how to reproduce evaluation results for the deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B model on the MATH-500 dataset, including specific server launch commands, evaluation commands, and the resulting Pass@1 scores across different difficulty levels and prompt engineering variations.
• Refined Sampling Logic for TPUs: The sampling logic within python/sgl_jax/srt/layers/sampler.py has been updated for TPU environments. A new function, top_k_top_p_min_p_sampling_from_probs_jax_tpu_runtime_sort_version, was introduced to explicitly sort probabilities before applying top-k, top-p, and min-p filtering, aiming for more precise control over the sampling process on TPUs.
• Simplified Conditional Code Paths: Conditional is_tpu_runtime() checks were removed from the multinomial and _apply_min_p_filter functions in sampler.py. This change standardizes the behavior of these functions, removing environment-specific branching for certain sampling operations.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[jimoosciuc]

LGTM