What are the main failure modes of LLM agents?
hardAnswer
- (1) Infinite loops: repeated tool calls with slight variations, spending tokens forever.
- Fix: , loop detection, cost caps.
- (2) Wrong tool selection: LLM chooses a tool that can't solve the problem.
- Fix: better tool descriptions, tool-selection few-shot.
- (3) Bad arguments: hallucinated JSON, missing fields.
- Fix: structured output / JSON schema.
- (4) Ignoring tool errors: model plows through on stale info.
- Fix: explicit error handling in the prompt.
- (5) Reward-hacking: model finds shortcuts .
- Fix: verification steps.
Check yourself — multiple choice
- Agents always succeed
- Infinite loops, wrong tool, bad args, ignored errors, reward-hacking — each needs specific safeguards (limits, schemas, errors, verification)
- Only latency issues
- Solved by fine-tuning
Agent failure modes: loops / wrong tool / bad args / ignored errors / hacking — engineer each safeguard.
#agents#reliability
Practise LLMs & GenAI
214 interview questions in this topic.
Related questions
- Why do LLMs hallucinate and how do you reduce hallucinations?
- What is sycophancy in LLMs and how do you reduce it?
- What is 'honesty' as an alignment target and how do you train for it?
- How does function calling / tool use work in modern LLMs?
- What is prompt chaining and why prefer it over one big prompt?
- How do you make an LLM cite its sources reliably?