AI & Language Infrastructure

Hinglish-Bench

Reference-Free Benchmark for Hinglish LLM Generation

34 prompts, 8 categories, CMI/M-Index/SyMCoM metrics with LLM judge. lm-evaluation-harness integration for evaluating Hindi-English code-mixed text generation.

CodeAll Projects
PythonNLPBenchmarkingLanguage Evaluationlm-eval-harness

Project Facts

NameHinglish-Bench
CategoryAI & Language Infrastructure
Primary LanguagePython
TechnologiesPython, NLP, Benchmarking, Language Evaluation, lm-eval-harness
Sourcehttps://github.com/instax-dutta/hinglish-bench

More in AI & Language Infrastructure

AgentLoop

Self-Verifying Autonomy Wrapper for Coding Agents

PhishScout

131 KB On-Device Phishing URL Detector

CL-Chat Reborn

P2P Encrypted Command-Line Chat

The Watcher

Reddit Used-Tech Intelligence Bot

PacketBuddy

Cross-Platform Network Usage Monitor

Roadmap Tutor

Learn Any roadmap.sh Roadline One Topic at a Time

What is Hinglish-Bench?

Hinglish-Bench is an open-source project by Sai Dutta Abhishek Dash, a founder and engineer based in Odisha, India. It belongs to the AI & Language Infrastructure category, and is a Reference-Free Benchmark for Hinglish LLM Generation. 34 prompts, 8 categories, CMI/M-Index/SyMCoM metrics with LLM judge. lm-evaluation-harness integration for evaluating Hindi-English code-mixed text generation. The primary implementation language is Python. It uses Python, NLP, Benchmarking, Language Evaluation, lm-eval-harness. Scores generations with CMI, M-Index, and SyMCoM, plus an LLM judge. Integrates with lm-evaluation-harness for standardized evaluation pipelines. Other systems in the same track are ornith-flight, AgentLoop, PhishScout, and MarkItDownJS. The full source code is public at https://github.com/instax-dutta/hinglish-bench. Sai Dutta Abhishek Dash has shipped 22 open-source production systems, maintains 85+ public repositories on GitHub, and has published 46+ models and 22+ datasets on Hugging Face. He is available for advisory roles and technical collaborations via contact@sdad.pro.