The gradient signal in teacher forcing is clean because the model is evaluated against the correct answer given the correct context at every step, producing a well-defined and unambiguous learning signal.

Authors

• Person: M. Brenndoerfer Organization: mbrenndoerfer.com
  Hallucination Causes: Why Language Models Fabricate Facts

Sources

• Hallucination Causes: Why Language Models Fabricate Facts mbrenndoerfer.com M. Brenndoerfer · mbrenndoerfer.com via serper

Referenced by nodes (1)

• Large Language Models concept