Evaluating agents

Where it is assessed

6 assessed items, 0 done.

• AIE-104 M7.3 Task completion, not transcript plausibility · tier 1
• AIE-104 M7.4 The plausible failure · tier 2
• AIE-104 M7.5 Retention: agent evaluation and state-based scoring · tier 1
• AIE-104 M9.2 The reliability report · tier 2
• AIE-104 M9.3 Publish · tier 3
• AIE-104 M9.4 Oral defense · tier 4

Verified sources

• Agentic AI MOOC Fall 2025 · Berkeley RDI
• LLM Agents MOOC Fall 2024 · Berkeley RDI
• Welcome to the Agents Course
• SWE-bench: Can Language Models Resolve Real-World GitHub Issues?
• tau-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains
• WebArena: A Realistic Web Environment for Building Autonomous Agents
• Why Do Multi-Agent LLM Systems Fail?
• UKGovernmentBEIS/inspect_ai