Gids-gekoppelde quiz · Middelmatig Niveau
Omkeringsvloek in de LLM-quiz
Test directionele feitenherinnering, omgekeerde vragen, trainingsvolgorde, experimentele controles en grenzen van de bevinding van de omkeervloek.
Vraag 1 van 8
Een model leerde ‘A is B’, maar slaagt er niet in te antwoorden ‘B is A’. Hoe heet dit?
Bekijk elke vraag8 vragen
Antwoorden en uitleg verschijnen terwijl je de quiz doet.
- Een model leerde ‘A is B’, maar slaagt er niet in te antwoorden ‘B is A’. Hoe heet dit?
- The lost-in-the-middle effect
- Catastrophic forgetting
- The reversal curse
- Modus ineenstorting
- Wat laat de omkeervloek zien over directionele associaties?
- Learning one word order does not guarantee reliable recall in the reverse order
- Language models automatically invert every relation
- A fact is always stored as a symmetric database edge
- Fine-tuning removes all directional errors
- Is de omkeringsvloek eenvoudigweg een mislukking omdat het model dit feit nooit heeft gezien?
- Yes, the model must not have read it
- It is caused only by a short context window
- It occurs only when facts are false
- No; it can fail in reverse even after training on the forward statement
- Waarom kan next-token-training bijdragen aan dit effect?
- It encodes all facts into an unordered table
- It trains on token order, so A-followed-by-B does not automatically teach B-followed-by-A
- It optimizes reverse queries by default
- It prevents memorization of entity names
- How can the effect be tested?
- Compare model performance on matched forward and reverse versions of facts
- Measure only the forward statement
- Change both the entities and question at once
- Use only a general language benchmark
- Wat impliceert het mislukken van een omgekeerde query over modelkennis?
- The model has no information about either entity
- The model is necessarily incapable of reasoning
- The model’s learned association may not support flexible inversion
- Every entity relation is symmetric
- Why is a reverse-direction test useful for a factual QA system?
- All questions use the same word order
- Users may ask for either endpoint of a relation
- It measures image classification accuracy
- It guarantees factual truth
- What should be inferred from one model’s reversal-cursed answer?
- All language models fail on every reversed fact
- The model has no semantic representations
- The relation must be false
- A documented failure mode to test, not proof that every model or fact behaves identically