b:{"metadata":[["$","title","0",{"children":"IndicSafe | Dataset | CivicDataSpace"}],["$","meta","1",{"name":"description","content":"

IndicSafe

A Benchmark for Evaluating Multilingual LLM Safety in South Asia

(ARR 2025)

📘 Overview

IndicSafe is a multilingual safety and alignment benchmark designed to evaluate large language models across major Indic languages.

It includes diverse prompts covering safety-critical categories such as:

Hate speech
Harassment
Misinformation
Self-harm and violence
Sensitive or unethical content

"}],["$","meta","2",{"name":"keywords","content":"LLM Safety"}],["$","meta","3",{"property":"og:title","content":"IndicSafe"}],["$","meta","4",{"property":"og:description","content":"

IndicSafe

A Benchmark for Evaluating Multilingual LLM Safety in South Asia

(ARR 2025)

📘 Overview

IndicSafe is a multilingual safety and alignment benchmark designed to evaluate large language models across major Indic languages.

It includes diverse prompts covering safety-critical categories such as:

Hate speech
Harassment
Misinformation
Self-harm and violence
Sensitive or unethical content

"}],["$","meta","5",{"property":"og:url","content":"https://dev.civicdataspace.in/datasets/ea5ad46b-2f33-46c5-a8a9-07611a430448"}],["$","meta","6",{"property":"og:site_name","content":"CivicDataSpace"}],["$","meta","7",{"property":"og:locale","content":"en_US"}],["$","meta","8",{"property":"og:image","content":"https://dev.civicdataspace.in/og.png"}],["$","meta","9",{"property":"og:type","content":"website"}],["$","meta","10",{"name":"twitter:card","content":"summary_large_image"}],["$","meta","11",{"name":"twitter:creator","content":"CivicDataLab"}],["$","meta","12",{"name":"twitter:title","content":"IndicSafe"}],["$","meta","13",{"name":"twitter:description","content":"

IndicSafe

A Benchmark for Evaluating Multilingual LLM Safety in South Asia

(ARR 2025)

📘 Overview

IndicSafe is a multilingual safety and alignment benchmark designed to evaluate large language models across major Indic languages.

It includes diverse prompts covering safety-critical categories such as:

Hate speech
Harassment
Misinformation
Self-harm and violence
Sensitive or unethical content

"}],["$","meta","14",{"name":"twitter:image","content":"https://dev.civicdataspace.in/og.png"}]],"error":null,"digest":"$undefined"}