References: Prompt Engineering for Token Efficiency
-
Prompt engineering - Wikipedia - Comprehensive coverage of prompt design patterns including zero-shot, few-shot, chain-of-thought, and instruction-tuning that this chapter applies through a token-cost lens.
-
In-context learning (natural language processing) - Wikipedia - Coverage of how few-shot examples shape model behavior at inference time, relevant to this chapter's few-shot pruning discussion.
-
Chain-of-thought prompting - Wikipedia - Coverage of the reasoning-step-elicitation pattern that the chapter's reasoning-budget discussion regulates.
-
AI Engineering - Chip Huyen - O'Reilly - The prompt engineering chapters give a vendor-neutral foundation that this chapter extends with token-cost analysis; particularly useful for the system-prompt-hygiene material.
-
Hands-On Large Language Models - Jay Alammar and Maarten Grootendorst - O'Reilly - The prompting chapters contain side-by-side before/after examples that illustrate the token-reduction techniques discussed here.
-
Anthropic Prompt Engineering Guide - Anthropic - Authoritative vendor reference covering system prompts, examples, XML tags, and chain-of-thought patterns; the recommended Anthropic-specific complement to this chapter.
-
OpenAI Prompt Engineering Guide - OpenAI - Vendor reference for prompting GPT and o-series models including instruction-following techniques specific to OpenAI training.
-
Google Gemini Prompt Design Strategies - Google - Vendor reference for prompting Gemini including its multimodal and long-context-specific patterns.
-
Prompt Engineering Guide (promptingguide.ai) - Elvis Saravia / DAIR.AI - Comprehensive open-source guide covering 50+ prompting techniques with examples; ideal companion for engineers who want a broader survey beyond the cost-focused treatment in this chapter.
-
Lilian Weng: Prompt Engineering - Lilian Weng - Long-form blog post categorizing prompting techniques by mechanism; the framework helps reason about which techniques add tokens and which remove them.