Crossword
Across
- 5. Safety measure bypass
- 9. LLM's apt analogy
- 10. LLM's core task
- 12. Malicious training data
- 15. Fine-tuning data type
- 16. Large language models
- 17. Training cost (USD)
Down
- 1. LLM's learned essence
- 2. Code using weights
- 3. Model's weights file
- 4. Challenge for self-improvement
- 6. Instruction hijacking
- 7. Helpful assistant creation
- 8. Fast, intuitive thinking
- 11. LLM's evolving capability
- 13. Training data size
- 14. Slow, deliberate thinking