Case Study

Analyzing a Language Model's Reasoning Process

A language model is prompted to solve the following word problem and show its work: 'Tom starts with 12 marbles. He then wins 7, loses 5, and is given another 3. How many marbles does Tom have in the end?' The model provides the following reasoning: 'First, I will calculate the net change from winning and losing marbles, which is 7 - 5 = 2. Tom started with 12 marbles, so adding this net change gives 12 + 2 = 14. Finally, he is given another 3, so 14 + 3 = 17. The final answer is 17.' Although the final answer is correct, analyze the model's reasoning process. Identify the primary flaw in its method when compared to the narrative sequence of the problem, and explain why this flawed approach could lead to incorrect answers in other, more complex scenarios.

0

1

Updated 2025-10-04

Contributors are:

Who are from:

Tags

Ch.2 Generative Models - Foundations of Large Language Models

Foundations of Large Language Models

Computing Sciences

Foundations of Large Language Models Course

Analysis in Bloom's Taxonomy

Cognitive Psychology

Psychology

Social Science

Empirical Science

Science