Concept List: Token Optimization

This is the enumerated concept list for the Token Optimization course. Each concept has a unique ConceptID (1-470) used throughout the learning graph.

Generative AI
Large Language Model
Foundation Model
Transformer Architecture
Autoregressive Generation
Token
Input Token
Output Token
Cached Token
Reasoning Token
Token Count
Tokenizer
Byte Pair Encoding
SentencePiece
Vocabulary Size
Context Window
Context Length
Prompt
System Prompt
User Message
Assistant Message
Conversation Turn
Multi-Turn Dialogue
Streaming Response
Stop Sequence
Max Tokens Parameter
Temperature
Top P Sampling
Logprobs
Model Family
Subword Tokenization
BPE Merge Rules
Tiktoken Library
Claude Tokenizer
Gemini Tokenizer
Token-To-Char Ratio
Whitespace Handling
Unicode Normalization
Special Tokens
End-Of-Sequence Token
Beginning-Of-Sequence Token
Padding Token
Token Boundary
Pre-Tokenization
Cross-Vendor Tokenizer Drift
Token Counting API
Local Token Estimation
Token Count Caching
Code Tokenization
Multilingual Tokenization
Per-Million-Token Price
Input Token Price
Output Token Price
Cached Input Price
Output Premium
Unit Economics
Cost Per Request
Cost Per Feature
Cost Per User
Cost Per Outcome
Cost Attribution
Token Budget
Monthly Token Spend
Forecasting Token Cost
Cost-Quality Tradeoff
Cost-Latency Tradeoff
Pareto Frontier
Pricing Tier
Volume Discount
Batch Discount
Enterprise Pricing
Free Tier Limit
Rate Limit
Quota Management
Burn Rate
Anthropic API
Claude Messages API
Claude Model Family
Claude Opus
Claude Sonnet
Claude Haiku
Anthropic SDK
API Key Management
Anthropic Prompt Caching
Cache Control Parameter
Cache Breakpoint
Cache TTL
Cache Read Tokens
Cache Write Tokens
Extended Thinking
Thinking Token Budget
Claude Tool Use
Tool Definition Schema
Tool Result Block
Message Content Block
Anthropic System Prompt
Stop Reason
Anthropic Streaming
Anthropic Batch API
Claude Vision Input
OpenAI API
Chat Completions API
OpenAI Responses API
OpenAI Model Family
GPT Model Series
OpenAI O Series
Reasoning Model
OpenAI SDK
Function Calling
Tool Choice Parameter
JSON Mode
Structured Outputs
Response Format
OpenAI Streaming
OpenAI Batch API
OpenAI Embeddings
OpenAI Fine Tuning
OpenAI Vision
Logit Bias
Seed Parameter
Token Usage Object
Prompt Tokens Field
Completion Tokens Field
Total Tokens Field
OpenAI Rate Limits
Google Gemini API
Gemini Model Family
Gemini Pro
Gemini Flash
Gemini Ultra
Gemini SDK
Long Context Window
One Million Context
Gemini Function Calling
Gemini Tool Config
Gemini Streaming
Gemini Batch Mode
Gemini Caching
Vertex AI
Google AI Studio
Gemini Safety Settings
Gemini Multimodal Input
Gemini Code Execution
Gemini Grounding
Gemini Token Counting
AI Coding Harness
Agentic Loop
Tool Use Loop
Claude Code
Claude Code Session
Claude Code Hooks
OpenAI Codex CLI
Codex Session
Google Antigravity
Antigravity Workspace
Harness System Prompt
Harness Token Overhead
Session Token Accumulation
Per-Session Token Cost
Conversation Compaction
Auto Compaction
Manual Compaction
Tool Call Iteration
Multi-Step Reasoning
Subagent Pattern
Agent Memory
Persistent Memory File
Working Directory Context
File Read Tool
File Edit Tool
Skill
Skill Description
Skill Body
Skill Trigger
Skill Invocation
Skill Frontmatter
Skill Bundle
Bundled Script
Skill Asset File
Lazy Skill Loading
Eager Skill Listing
Task Decomposition
Task-Skill Binding
Skill Selection
Skill Misfire
False Positive Trigger
False Negative Trigger
Trigger Precision
Skill Library
Anthropic Skill Format
Script Delegation
Shell Script Skill
Python Script Skill
Skill Refactoring
Token Reduction Ratio
Structured Logging
Log Schema Design
Log Line
JSON Log Format
Log Field
Required Log Field
Optional Log Field
Model Field
Prompt Hash
Input Token Field
Output Token Field
Cached Token Field
Latency Field
Cost Field
Feature Tag
User Identifier
Outcome Field
Trace Identifier
Span Identifier
Request Identifier
Session Identifier
PII Redaction
Prompt Truncation In Logs
Log Sampling
Log Retention Policy
Observability
OpenTelemetry
OTel LLM Conventions
Metric
Counter Metric
Histogram Metric
Dashboard
Cost Dashboard
Hit Rate Dashboard
Latency Dashboard
Token Volume Chart
Time Series Aggregation
Anomaly Detection
Alerting Rule
Token Spike Alert
Cost Threshold Alert
Cardinality Concern
Aggregation Period
Drill-Down Analysis
Cross-Service Tracing
Log File Analysis
Log Aggregation
Top-N Cost Drivers
Pareto Analysis
Outlier Detection
Runaway Prompt
Pathological Agent Loop
Cost Hotspot
Per-Feature Cost Roll-Up
Per-User Cost Roll-Up
Per-Model Cost Roll-Up
Prompt Template Grouping
Cohort Analysis
Funnel Analysis
Histogram Of Token Counts
P50 Token Usage
P95 Token Usage
P99 Token Usage
Long-Tail Cost
Analysis Notebook
A/B Testing
Hypothesis
Null Hypothesis
Control Group
Treatment Group
Traffic Split
Random Assignment
Stratified Assignment
Primary Metric
Guardrail Metric
Quality Metric
Cost Metric
Latency Metric
Satisfaction Metric
Sample Size Calculation
Statistical Power
Statistical Significance
P-Value
Confidence Interval
Effect Size
Stopping Rule
Sequential Testing
Multi-Armed Bandit
CUPED Adjustment
Novelty Effect
Prompt Engineering
System Prompt Hygiene
Instruction Compression
Few-Shot Example
Few-Shot Pruning
Zero-Shot Prompting
Chain Of Thought
Dead Context
Redundant Instruction
Verbose Boilerplate
Prompt Template
Template Versioning
Prompt Variable
Variable Interpolation
Prompt Compression Tool
Selective Compression
Prompt Length Budget
Output Length Budget
Concise Output Instruction
Token-Aware Rewriting
Whitespace Stripping
Comment Removal
Schema Minimization
Symbol Substitution
Reusable Prompt Block
Prompt Caching
Cache Key
Cache Hit
Cache Miss
Cache Hit Rate
Cache Warming
Cache Invalidation
Cache Invariant
Stable Prefix
Volatile Suffix
Cache Boundary
Cross-Vendor Caching
Cache Cost Savings
Cache Monitoring
Cache Hit Rate Metric
Cache Eviction
Implicit Caching
Explicit Caching
Cache Aware Routing
Cache Stampede
Retrieval Augmented Generation
Embedding
Vector Database
Chunking
Chunk Size
Chunk Overlap
Top-K Retrieval
Reranker
Cross-Encoder Reranker
Retrieval Score
Context Injection
Retrieved Context Bloat
Context Pruning
Hybrid Retrieval
BM25 Retrieval
Dense Retrieval
Query Rewriting
HyDE
Document Compression
Summarization-Based RAG
Citation Of Sources
RAG Cost Analysis
Context Quality Metric
Retrieval Precision
Retrieval Recall
Context Window Budget
Sliding Window
Conversation Summarization
Compaction Strategy
Hierarchical Summary
Memory File
Long-Term Memory
Short-Term Memory
Context Truncation
Pre-Send Token Counting
Context Quality Decay
Lost-In-The-Middle
Context Reordering
Selective Context Inclusion
Context Eviction Policy
Model Routing
Cheap-First Cascade
Escalation Trigger
Confidence Threshold
Quality Gate
Fallback Model
Cross-Vendor Routing
Task Classifier
Difficulty Estimation
Routing Policy
Routing Cost Savings
Routing Quality Risk
Per-Task Model Selection
Vendor Lock-In Risk
Vendor-Neutral Abstraction
Max Tokens Setting
Stop Sequence Setting
Length Penalty
JSON Schema Output
Concise Mode
Verbosity Parameter
Reasoning Budget
Thinking Token Limit
Truncation Detection
Streaming Cancellation
Early Stopping
Output Postprocessing
Output Validation
Schema Enforcement
Reasoning Effort Setting
Agent Budget Policy
Per-Session Token Budget
Per-Session Tool Call Budget
Loop Iteration Limit
Wall Clock Limit
Cost Cap
Graceful Degradation
Budget Exhaustion Handling
Runaway Detection
Circuit Breaker Pattern
Tool Call Throttling
Subtask Budget Allocation
Budget Audit Log
Budget Reporting
Per-Engineer Budget
Per-Repository Budget
Per-PR Budget
Budget Notification
Manager Weekly Report
Budget Versus Outcome
Batch API
Asynchronous API
Batch Job Submission
Batch Job Status
Batch Window
Batch Discount Rate
Batch Versus Synchronous
Throughput Optimization
Latency Tolerance
Job Queue
Result Polling
Webhook Notification
Idempotency Key
Retry Policy
Backoff Strategy
Data Privacy
PII Detection
Sensitive Field Redaction
Compliance Risk
GDPR
HIPAA
SOC2 Audit
Data Residency
Vendor Data Retention
Opt-Out Of Training
Logging Privacy Risk
Hashing Sensitive Strings
Tokenized Identifier
Audit Trail
Anonymization Strategy
Baseline Cost Measurement
Optimization Hypothesis
Quality Regression Detection
Before-After Report
Optimization Backlog
Cost Reduction Target
Pilot Rollout
Canary Deployment
Token Dashboard Project
Vendor-Neutral Logging Project
Skill Refactor Project
Budget Policy Document
Engineering Manager Review
Cost Reduction Postmortem
Reproducible Benchmark
Eval Suite
Golden Test Set
Regression Test Loop
Continuous Cost Monitoring
Token Efficiency Roadmap
5-Hour Limit
Weekly Limit
Serial Execution
Parallel Execution
Parallel Token Penalty