The Reversal Curse: LLMs trained on “A is B” fail to learn “B is A”

Our summary

Demonstrates that models which learn a fact in one direction often fail to generalize it to the reverse, indicating that learned associations are not stored as symmetric relations.

Why it matters

A concrete, reproducible limit on how LLMs represent knowledge — relevant to any factual-lookup or reasoning task.

Related findings (1)

• The reversal curse: 'A is B' not generalizing to 'B is A'Medium

  A model trained that 'A is B' frequently fails to answer 'B is ?', revealing that learned relations are not symmetric.

Cite this