Responsible AI: Assessing the Truthfulness of AI Model Responses
summary
Describes methods and tools to evaluate the truthfulness of answers produced by large language models using Ground Truth datasets, and shows how to use LLMs themselves to detect textual inconsistencies when Ground Truth is not available. Mentions practical evaluation workflows leveraging LLM capabilities and datasets, aimed at responsible AI verification.
AI-written summary · about
chapters (5)
- 0:00Responsible AI introduction
- 6:32Truthfulness problems in AI
- 12:39Evaluating with truth datasets
- 18:20Evaluating with another model
- 22:33Summary and closing
related talks
22mFrom the Trenches: Automated Testing of Large Language Models
29mLies, damned lies and large language models
13mNew Challenges in Software Quality Brought by GenAI and LLMs
25mAI vs. AI: How to Test AI Systems Using AI
46mCan you trust your AI?
51mTowards a Pragmatic Understanding of Large Language Models
26mResponsible AI-What, Why, How and Future!
46mEnhancing Agent Reliability Through Evaluation-Driven Development