AI Data Evaluator at DataAnnotation.tech
Conduct logical evaluation of LLM-generated text outputs to ensure factual correctness, accuracy, and adherence to detailed instructional guidelines. Perform adversarial testing and quality assurance to detect issues such as hallucinations and potential bias in model responses. Apply critical thinking to assess multi-step prompt handling and confirm the model provides helpful and safe responses. • Review and rate LLM outputs for accuracy and guideline compliance. • Run adversarial/QA checks to surface hallucinations and bias. • Validate that responses follow nuanced instructions and safety expectations. • Evaluate complex prompt-response behavior in a structured workflow.