EXPLORE / 06
Terms index
Specialized AI and software-development terms with concise explanations
RECORDS
Entity records
3375 records · 48/85
- 1881
Static Analysis
Method of checking source code or binaries for vulnerabilities without executing them
Website pending - 1882
Model Security
Measures protecting ML models from tampering, reverse engineering, or adversarial attacks
Website pending - 1883
Long Video QA
Question answering tasks targeting videos with longer durations and complex temporal information
Website pending - 1884
Adversarial Dataset Refinement
A method using adversarial mechanisms or strong models to generate distractors for filtering and optimizing dataset quality
Website pending - 1885
Zero-configuration
Mechanism that automates deployment and optimization without manual parameter setup.
Website pending - 1886
Sampling Strategy
Methods determining how LLMs select the next token from probability distributions.
Website pending - 1887
Multilingual LLM
Large language models capable of processing multiple natural languages
Website pending - 1888
Assistant Model
A smaller model that provides preliminary predictions or drafts during the main model's inference process.
Website pending - 1889
Open Dataset
Machine learning data collections freely accessible, usable, and modifiable by the public
Website pending - 1890
Human Preference Alignment
Training process that aligns model behavior with human values and preferences.
Website pending - 1891
Early Exit
A strategy allowing data to yield results at shallower network layers, thereby reducing computation and accelerating inference.
Website pending - 1892
Self-Attention
Mechanism allowing parallel processing and focusing on relationships between parts
Website pending - 1893
Positional Encoding
Vector representation method injecting sequence order info into Transformers
Website pending - 1894
Multi-Agent Debate
A technique where multiple AI agents improve answer quality by questioning and arguing with each other.
Website pending - 1895
Data Labeling
The process of adding tags or annotations to raw data for supervised learning purposes.
Website pending - 1896
Named Entity Recognition (NER)
A technique that identifies and classifies named entities like persons, locations, and organizations in text.
Website pending - 1897
Pretrained Weights
The set of parameters obtained after training a model on large-scale datasets, used for fine-tuning downstream tasks
Website pending - 1898
SigLIP
An improved visual encoding architecture based on sigmoid loss function, derived from CLIP
Website pending - 1899
Model Sharding
Distributing large model parameters across multiple device memories to support training or inference.
Website pending - 1900
JAX
A Python library for high-performance numerical computing and machine learning.
Website pending - 1901
LLM-as-a-Judge
Method using large language models to evaluate the quality of generated outputs.
Website pending - 1902
Decoder-only
A model architecture using only the Transformer decoder stack for autoregressive generation.
Website pending - 1903
Flash Attention
An algorithm that accelerates attention computation and reduces memory usage by optimizing memory access patterns.
Website pending - 1904
SwiGLU
An efficient feed-forward network structure combining Swish activation with Gated Linear Units.
Website pending - 1905
RoPE
A method that incorporates absolute position information into relative positions within attention mechanisms.
Website pending - 1906
Inference Efficiency
Performance of a model in terms of computation speed and resource consumption during output generation
Website pending - 1907
State Space Model (SSM)
An efficient architecture for sequence modeling characterized by linear complexity
Website pending - 1908
Logits
Unnormalized raw scores from the output layer, converted to probabilities via Softmax
Website pending - 1909
Nucleus Sampling
Random sampling from a candidate set whose cumulative probability exceeds a threshold
Website pending - 1910
Greedy Search
Decoding strategy selecting the highest probability token at each step, prone to repetition
Website pending - 1911
Memory Visualization
Graphical representation of GPU memory allocation, deallocation, and fragmentation
Website pending - 1912
CUDA OOM
Runtime error caused by insufficient GPU memory during deep learning training
Website pending - 1913
Multi-task Learning
Training a single model to optimize multiple related tasks simultaneously to improve generalization.
Website pending - 1914
Static Embeddings
Pre-computed word vector representations that do not require dynamic generation at runtime.
Website pending - 1915
torch.compile
Just-in-time compilation technology in PyTorch 2.0 for optimizing computational graph performance.
Website pending - 1916
Open Neural Network Exchange
An open standard format for representing deep learning models, enabling cross-platform deployment
Website pending - 1917
Inference Infrastructure
Hardware and software systems supporting LLM execution, request handling, and result return
Website pending - 1918
Reasoning Enhancement
Improving logical deduction and problem-solving by increasing compute resources or optimizing algorithms
Website pending - 1919
Action Tokenization
Technique converting continuous robot control signals into discrete sequences for Transformer processing.
Website pending - 1920
Open Source Reproduction
Retraining on public code and data to validate or mimic the performance of specific closed-source models.
Website pending