As a Remote AI Trainer & Data Evaluator, I assess large language model responses for truthfulness, helpfulness, safety, and instruction-following. I write prompts to test reasoning and coding, then use web research to fact-check claims and document why responses meet project guidelines.
On the LLM Evaluation & Fact-Verification Benchmark, I evaluated 300+ model responses and documented error patterns to assist engineers with dataset refinement. I also developed Python scripts to clean and reformat datasets into JSON and CSV for evaluation workflows.
