Today for AI

EXPLORE / 06

Terms index

Specialized AI and software-development terms with concise explanations

Entity records

4517 records · 78/113

  1. 3081

    Agentic Automation

    Use of software that reasons and adapts to autonomously complete tasks

    Website pending
  2. 3082

    Human-AI Trust

    The subjective confidence users place in an AI system's reliability, safety, and intent.

    Website pending
  3. 3083

    Compute Capacity

    Maximum computational processing capability provided by a data center

    Website pending
  4. 3084

    AI Cloud Service

    Cloud infrastructure optimized specifically for AI workloads

    Website pending
  5. 3085

    Request for Information (RFI)

    A process by which government agencies collect opinions and data from the public or stakeholders before formulating policy.

    Website pending
  6. 3086

    Open AI / Open Source AI

    AI systems where code, data, or model weights are publicly available for free use, modification, and distribution.

    Website pending
  7. 3087

    LiveCodeBench

    A benchmark evaluating code generation models' real-time programming capabilities

    Website pending
  8. 3088

    Local LLM

    Language models run on personal devices rather than cloud servers

    Website pending
  9. 3089

    Content-Addressable Storage

    A storage method that uses data content hashes instead of paths to locate and store data

    Website pending
  10. 3090

    Base Model

    Pre-trained large model with general capabilities, not fine-tuned for specific tasks

    Website pending
  11. 3091

    Scheduler

    Component deciding when to assign queued requests to compute resources

    Website pending
  12. 3092

    Request Queue

    Data structure storing pending inference requests to smooth traffic peaks

    Website pending
  13. 3093

    Web UI

    Browser-based graphical user interface simplifying interaction with complex software or models.

    Website pending
  14. 3094

    LLM Leaderboard

    A system that ranks multiple large language models based on performance in specific benchmarks.

    Website pending
  15. 3095

    Real-Time Communication (RTC)

    Technologies enabling low-latency bidirectional audio, video, and data transfer

    Website pending
  16. 3096

    Supply Chain Security

    Measures protecting software and data from tampering and attacks throughout the lifecycle from development to deployment

    Website pending
  17. 3097

    Model Scanning

    The process of analyzing AI model files statically or dynamically to detect vulnerabilities, backdoors, or malicious code

    Website pending
  18. 3098

    UI

    The visual and interactive layer for displaying data and controlling functions

    Website pending
  19. 3099

    Decode

    The autoregressive process of generating output tokens one by one based on context

    Website pending
  20. 3100

    Prefill

    The phase where an LLM processes input prompts and generates initial KV cache

    Website pending
  21. 3101

    Pipeline Parallelism

    A distributed training strategy that splits model layers across devices for sequential execution

    Website pending
  22. 3102

    Chat Template

    Structured text framework defining dialogue format and special tokens

    Website pending
  23. 3103

    Content Safety Classifier

    A specialized model designed to detect and intercept harmful, violating, or sensitive content

    Website pending
  24. 3104

    LeRobot

    An open-source library and dataset platform for robot learning by Hugging Face

    Website pending
  25. 3105

    PyTorch

    Widely used open-source deep learning framework supporting dynamic computation graphs

    Website pending
  26. 3106

    1.58-bit Quantization

    Extreme quantization technique compressing weights to ~1.58 bits

    Website pending
  27. 3107

    Hybrid-Head Architecture

    A model structure combining different attention mechanisms or expert modules to optimize computational efficiency

    Website pending
  28. 3108

    TRL

    An open-source library for fine-tuning and RLHF of LLMs

    Website pending
  29. 3109

    Kernel Fusion

    Combining multiple GPU operations to reduce memory access overhead

    Website pending
  30. 3110

    LLM Inference Engine

    Software systems designed for efficient deployment and execution of LLMs

    Website pending
  31. 3111

    Robot Foundation Model

    Large-scale pre-trained general models adaptable to various robot tasks and hardware

    Website pending
  32. 3112

    Evaluation Suite

    A collection of standardized tasks and metrics used to systematically test model performance

    Website pending
  33. 3113

    LPU

    Deterministic hardware architecture designed by Groq specifically for LLM inference.

    Website pending
  34. 3114

    Language Processing Unit

    Hardware architecture by Groq designed specifically to accelerate language model inference

    Website pending
  35. 3115

    Backend Abstraction

    Architectural pattern that hides differences in underlying hardware or libraries behind a unified interface

    Website pending
  36. 3116

    Backend Integration

    Process of connecting different software components to work together

    Website pending
  37. 3117

    Finetuning

    Continuing training on a pretrained model for specific tasks

    Website pending
  38. 3118

    Early Training Evaluation

    Methods to monitor and analyze model performance or convergence trends during the initial stages of training

    Website pending
  39. 3119

    Sparse Embedding

    A high-dimensional vector representation where most dimensions are zero, used for efficient retrieval

    Website pending
  40. 3120

    Open-source Hardware

    Hardware designs released under licenses allowing free modification and distribution

    Website pending