Enabling small language models to solve complex reasoning tasks
December 12, 2025
December 12, 2025
CAMBRIDGE, Massachusetts, Dec. 12 -- The Massachusetts Institute of Technology posted the following news:
* * *
Enabling small language models to solve complex reasoning tasks
*
As language models (LMs) improve at tasks like image generation, trivia questions, and simple math, you might think that human-like reasoning is around the corner. In reality, they still trail us by a wide margin on complex tasks. Try playing Sudoku with one, for instan . . .
* * *
Enabling small language models to solve complex reasoning tasks
*
As language models (LMs) improve at tasks like image generation, trivia questions, and simple math, you might think that human-like reasoning is around the corner. In reality, they still trail us by a wide margin on complex tasks. Try playing Sudoku with one, for instan . . .