Evaluation short description

• Why is this evaluation interesting?
  ArmBench-LLM adds standardized Armenian-language evaluation across 24 tasks, including classification, QA, reasoning, summarization, translation, NER, and language correction. Armenian currently has very limited coverage in common LLM evaluation frameworks.

• How used is it in the community?
  The benchmark was released recently, so adoption is still growing. Its dataset, results, and evaluation code are publicly available, and the proposed LightEval integration is implemented in PR Add Armenian language evaluation suite (ArmBench-LLM) #1289.

Evaluation metadata

• Paper/report URL: https://huggingface.co/blog/Metric-AI/armbench-llm
• GitHub/implementation URL: Add Armenian language evaluation suite (ArmBench-LLM) #1289
• Dataset URL: https://huggingface.co/datasets/Metric-AI/ArmBench-LLM-data

Related PR: #1289