Prompt Engineering
In-context learning, chain of thought, structured output, compression and injection-aware design.
10concepts
94flashcards
73minutes of reading
- 01 Automatic Prompt Optimisation Treating the prompt as a parameter to be searched rather than a string to be tweaked, using a metric, a dataset, and an LLM that proposes and critiques its own instructions.
- 02 Prompt Compression Cutting prompt tokens while holding task performance, via perplexity-based token dropping (LLMLingua) or learned gist tokens, and when prompt caching beats both.
- 03 Prompt Format Sensitivity Semantically identical prompts that differ only in separators, spacing, or option order can move accuracy by tens of points, which makes any single-format benchmark number a sample rather than a measurement.