Techniques.

Beginner · foundation/honesty

Ask before inventing

Require the model to request missing critical inputs instead of fabricating.

When

Facts, numbers, legal/medical, or branded claims.

AnyClaudeGPTGrokGeminiLlama

Expert · foundation/prompt-engineering/optimization

BayesPO: Bayesian Prompt Optimization

BayesPO optimizes prompts by treating them as discrete tokens in a Bayesian posterior sampling framework, balancing task performance with language model fluency.

When

When seeking to automatically optimize prompts for complex tasks or nuanced generation, especially when balancing task-specific rewards with general language model coherence.

AnyClaudeGPTGrokGeminiLlama

Beginner · foundation/persona

Biographical Persona Prompting

Using biographical personas in system prompts to influence LLM behavior, potentially affecting output length, style, and correctness depending on the model and persona.

When

When you want to explore how different 'identities' or roles assigned to the LLM might alter its response characteristics, such as verbosity or the inclusion of disclaimers.

ClaudeGPT

Expert · foundation/distillation/reasoning

Bootstrapped Iterative Self-Reasoning Distillation

Compress LLM reasoning chains by distilling concise, correct traces, then using them to guide further distillation for efficiency and accuracy.

When

When needing to reduce the length and computational cost of LLM reasoning chains without losing accuracy, especially for deployment on resource-constrained systems.

AnyClaudeGPTGrokGeminiLlama

Expert · foundation/evaluation/llm

CRAFT: LLM Capability Diagnosis

Diagnoses LLM weaknesses by clustering rubric criteria into a hierarchy, scoring performance at each level, and identifying specific failure points for targeted fine-tuning data generation.

When

When you have a rubric-based evaluation dataset and need to understand not just where your model fails, but why, in order to generate targeted data for fine-tuning and improve specific capabilities.

AnyClaudeGPTGrokGeminiLlama

Beginner · foundation/agentic-workflows

EpiNarrate Agentic Narrative Generation

An agentic framework for generating grounded public health narratives from complex epidemiological projections by separating numerical reasoning from natural language generation.

When

When needing to create clear, consistent, and quantitatively grounded narratives from multi-dimensional, ensemble-based numerical data, especially in public health or similar data-heavy communication contexts.

Any

Beginner · foundation/constraints

Hard constraints

Explicit boundaries the model must not violate.

When

When drift, hallucination, or tone risk is high.

AnyClaudeGPTGrokGeminiLlamaImage

Intermediate · foundation/agents

Knowledge-Centric Agents

A framework for workflow generation in visual creation systems that emphasizes modeling and reasoning with knowledge across multiple abstraction levels, rather than direct text-to-JSON generation.

When

When generating complex, structured outputs like visual creation workflows (e.g., ComfyUI) that require expert-level reasoning and hierarchical understanding, and where direct text-to-code generation is insufficient.

Any

Beginner · foundation/examples

Multishot examples

Show 1–3 miniature input→output pairs when the pattern is hard to describe.

When

Style transfer, classification, or unusual formats.

AnyClaudeGPTGrokGeminiLlama

Beginner · foundation/format

Output schema

Force a reusable structure the user can copy or parse.

When

Whenever the deliverable has parts (plan, draft, checklist).

AnyClaudeGPTGrokGeminiLlama

Expert · foundation/prompting/iterative-refinement

Recursive Harness Self-Improvement

Iteratively refine agent harness specifications using pairwise feedback from revision history to boost agent performance and data quality for training.

When

When optimizing an agent's task-specific harness to improve its performance and the quality of generated data for future model training, particularly for agents with low reasoning effort.

AnyClaudeGPTGrokGeminiLlama

Beginner · foundation/role

Role + objective

State who the model is and what success looks like before the task.

When

Almost every non-trivial prompt.

AnyClaudeGPTGrokGeminiLlama