← Back to AI Factory Chat

AI Factory Glossary

189 technical terms and definitions

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z All
Showing page 3 of 4 (189 entries)

token forcing, llm optimization

Token forcing mandates specific tokens at designated positions.

token importance scoring, llm architecture

Assign importance scores to determine computation allocation.

token limit in prompts, generative models

Maximum prompt length.

token merging, transformer

Combine similar tokens to reduce sequence length.

token streaming, llm optimization

Token streaming sends individual tokens immediately rather than waiting for completion.

token-to-parameter ratio, training

Training tokens per parameter.

tokenizer training, nlp

Create tokenizer vocabulary.

tool availability,production

Percentage of time tool is ready.

tool calling agent, ai agents

Tool calling agents invoke external functions APIs or resources to accomplish tasks.

tool calling with validation,ai agent

Validate tool call arguments before execution.

tool discovery, ai agents

Tool discovery enables agents to find and learn about available functions dynamically.

tool documentation, ai agents

Tool documentation describes function capabilities parameters and expected outputs for agent understanding.

tool idle management, environmental & sustainability

Tool idle management powers down unused equipment components reducing standby energy consumption.

tool matching maintenance, production

Keep tools similar.

tool result parsing, ai agents

Tool result parsing extracts relevant information from function outputs for agent reasoning.

tool selection, ai agents

Tool selection chooses appropriate functions from available repertoire for current needs.

tool use / function use,ai agent

LLM decides when and how to call external APIs tools or functions.

tool use training, fine-tuning

Teach models to use external tools.

tool-augmented llms,ai agent

Equip models with calculators search APIs code execution.

toolbench, ai agents

ToolBench evaluates agent ability to use diverse APIs and tools effectively.

toolformer,ai agent

Model trained to decide when and how to use tools.

top-k routing, llm architecture

Top-k routing selects k highest-scoring experts for each token.

topk pooling, graph neural networks

Top-K pooling selects a fixed number of highest-scoring nodes based on learned projection vectors for hierarchical graph representation.

topk pooling, graph neural networks

Select top-k nodes by importance.

topological qubits, quantum ai

Qubits protected by topology.

topology-aware training, distributed training

Consider network structure.

torchscript, model optimization

TorchScript creates serializable and optimizable representations of PyTorch models.

total cost ownership, supply chain & logistics

Total cost of ownership includes purchase price plus logistics inventory quality and risk costs.

total productive maintenance, tpm, production

Holistic maintenance approach.

toxicity classifier,ai safety

Model trained to detect harmful language.

toxicity detection models, ai safety

Identify toxic content.

toxicity detection, ai safety

Toxicity detection identifies offensive abusive or hateful language in text.

toxicity detection,ai safety

Classify text for hate speech offensive language or harmful content.

toxicity prediction, healthcare ai

Predict toxic effects of compounds.

tracin, explainable ai

Efficient influence computation.

trades, trades, ai safety

Balance accuracy and robustness.

trailing edge / mature node,industry

Older larger process nodes still used for cost-sensitive products.

trailing-edge node, business & strategy

Trailing-edge nodes are mature processes offering stability and cost advantages.

training compute budget, planning

Total FLOPS for training.

training cost estimation, planning

Predict computational cost.

training cost,model training

Total compute and time required to train a model.

training data attribution, interpretability

Training data attribution identifies which training examples most influenced specific predictions.

training data extraction attack,ai safety

Attempt to extract memorized training examples from model.

training data quality vs quantity, data quality

Tradeoff between data aspects.

training efficiency metrics, optimization

Measure computational efficiency.

training job orchestration, infrastructure

Manage complex training workflows.

training on thousands of gpus, distributed training

Massive-scale distributed training.

training pipeline optimization, optimization

Optimize end-to-end training workflow.

training time prediction, planning

Estimate training duration.

training verification, quality & reliability

Training verification confirms personnel understand and can execute procedures.