Error types of generative AI · Question 4 of 60
Why do LLMs easily make mistakes in multi-step arithmetic and logic tasks?
1 answer is correct. Easy Created with AI
Show answer
Answer
- B They continue learned text patterns instead of truly reasoning logically
Explanation
LLMs predict the likely next token. This often looks like logic but is pattern matching. In tasks with many dependent steps this quickly leads to wrong results.