Across
- 2. Deliberate reasoning
- 4. LLM's evolving capability
- 6. Source of knowledge
- 7. LLM's core function
- 10. Second training phase
- 12. Handling various data
- 13. Large language models
- 15. Instruction hijacking
Down
- 1. Malicious data
- 3. Fine-tuning data type
- 5. First training phase
- 8. New OS analogy
- 9. Model's knowledge
- 11. Safety measure bypass
- 14. Code that uses it