AI Cheat Sheets

One-page AI cheat sheets for developers: how each mechanism works, the rules that hold, and when to use it. Click a sheet to watch it build step by step.

  • Home /
  • AI Cheat Sheets

MMLU Benchmark cheat sheet

A high score shows what a model absorbed in training, not how it handles your own task.

#Model EvaluationAnimated

→ Topic: MMLU Benchmark

Model Context Protocol cheat sheet

MCP standardizes what an assistant can reach; vetting servers and adding auth stay your job.

#Agentic & Autonomous CodingAnimated

→ Topic: Model Context Protocol

Model Evaluation cheat sheet

No benchmark, metric or leaderboard settles fit alone; decide the metric first and test on your own tasks.

#Model EvaluationAnimated

→ Topic: Model Evaluation

Model Registry cheat sheet

A registry gives every model a version and a lifecycle state, so rollback is a pointer change, not a redeploy.

#LLMOps & ProductionAnimated

→ Topic: Model Registry

Model Routing cheat sheet

The router decides per request which model answers, so the decision needs a quality check, not only a cost rule.

#LLMOps & ProductionAnimated

→ Topic: Model Routing

Multi-Agent Systems cheat sheet

More agents add coordination, not intelligence; split the work only when one agent visibly fails.

#AI Agent ArchitectureAnimated

→ Topic: Multi-Agent Systems

Multi-Turn Prompt Design cheat sheet

The model keeps no memory between turns, so put constraints up front and reset a thread once it drifts.

#Prompt EngineeringAnimated

→ Topic: Multi-Turn Prompt Design

Multi-Vector Retrieval cheat sheet

Keeping one vector per token recovers nuance a single embedding averages away, but the index grows accordingly.

#Retrieval-Augmented GenerationAnimated

→ Topic: Multi-Vector Retrieval

Multimodal Architecture cheat sheet

Every input is projected into the token space the language model already reads, so lost detail starts at the connector.

#Transformer & Attention InternalsAnimated

→ Topic: Multimodal Architecture

Multimodal Prompting cheat sheet

Attach the visual and tell the model where to look; omission and fabrication come from the architecture, not the wording.

#Prompt EngineeringAnimated

→ Topic: Multimodal Prompting

Multimodal RAG cheat sheet

When the answer lives in a chart or table, retrieve the page itself, not a text transcript of it.

#Knowledge Retrieval SystemsAnimated

→ Topic: Multimodal RAG

Neural Network Basics for LLMs cheat sheet

A network is not programmed with rules: it repeats predict, measure, correct until its weights fit the training data.

#Neural Network ArchitecturesAnimated

→ Topic: Neural Network Basics for LLMs

Pre-Training cheat sheet

Pre-training decides what a model knows; every later stage only changes how it uses that knowledge.

#LLM Training & Pre-TrainingAnimated

→ Topic: Pre-Training

Precision Recall and F1 Score cheat sheet

Raising the threshold buys precision with missed positives; decide which error costs more before you tune.

#Model EvaluationAnimated

→ Topic: Precision Recall and F1 Score

Prompt Chaining cheat sheet

A chain is only as reliable as its weakest hand-off, so define each step's output before writing its prompt.

#Prompt EngineeringAnimated

→ Topic: Prompt Chaining

Prompt Engineering cheat sheet

A prompt is a stack of layers, and most failures trace back to a single mishandled layer.

#Prompt EngineeringAnimated

→ Topic: Prompt Engineering

Prompt Engineering for Image Generation cheat sheet

A prompt steers what the model already learned; pin the seed and model version before judging any wording change.

#AI Image Generation & EditingAnimated

→ Topic: Prompt Engineering for Image Generation

Prompt Injection cheat sheet

Any defense written inside the prompt can be overwritten by the prompt, so the safeguards that hold live outside the model.

#Prompt Ops & SecurityAnimated

→ Topic: Prompt Injection

Prompt Optimization cheat sheet

Define what better means first, change one clause at a time, and measure each change against that baseline.

#Prompt Ops & SecurityAnimated

→ Topic: Prompt Optimization

Prompt Testing and Evaluation cheat sheet

Offline runs catch regressions before deployment, online scoring catches the long tail, and each live failure joins the golden set.

#Prompt Ops & SecurityAnimated

→ Topic: Prompt Testing and Evaluation

Prompt Versioning and Management cheat sheet

Each version keeps its model settings and eval results, and moving the live label back undoes a bad change without a redeploy.

#Prompt Ops & SecurityAnimated

→ Topic: Prompt Versioning and Management

Quantization cheat sheet

Fewer bits barely move the average, but the loss lands unevenly, so test your own task.

#Inference OptimizationAnimated

→ Topic: Quantization

Query Transformation cheat sheet

Each rewrite technique fixes one kind of retrieval failure, and every rewrite adds a call that can drift or invent.

#Retrieval-Augmented GenerationAnimated

→ Topic: Query Transformation

RAG Evaluation cheat sheet

A wrong answer has two possible causes, so a faithful answer built on the wrong chunks still needs a retrieval fix.

#Retrieval-Augmented GenerationAnimated

→ Topic: RAG Evaluation

RAG Guardrails and Grounding cheat sheet

A guardrail catches an unsupported claim before the user reads it, but it does not make the model itself more accurate.

#Retrieval-Augmented GenerationAnimated

→ Topic: RAG Guardrails and Grounding

ReAct Prompting cheat sheet

The model's own reasoning picks each tool call, so every loop needs a clear stop condition and a narrow tool set.

#Prompt EngineeringAnimated

→ Topic: ReAct Prompting

Real-Time AI Generation cheat sheet

Step distillation shrinks the model's work, and warm streaming infrastructure removes the wait before it even starts.

#Generative Media Pipelines & APIsAnimated

→ Topic: Real-Time AI Generation

Recurrent Neural Network cheat sheet

One hidden state carries the whole past forward, which is why gates had to decide what survives.

#Neural Network ArchitecturesAnimated

→ Topic: Recurrent Neural Network

Red Teaming for AI cheat sheet

Every attack that lands becomes a guardrail, and the cycle starts again after each model or prompt change.

#Safety & Red TeamingAnimated

→ Topic: Red Teaming for AI

Reranking cheat sheet

A reranker sharpens precision on what retrieval already found; it cannot rescue a document that never made the list.

#Retrieval-Augmented GenerationAnimated

→ Topic: Reranking