Results for "high-risk"

AI predicting crime patterns (highly controversial).

AI-driven buying/selling of financial assets.

Returns above benchmark.

AI reinforcing market trends.

Isolating AI systems.

Signals indicating dangerous behavior.

Accelerating safety relative to capabilities.

A formal privacy framework ensuring outputs do not reveal much about any single individual’s data contribution.

A measure of a model class’s expressive capacity based on its ability to shatter datasets.

Measures a model’s ability to fit random noise; used to bound generalization error.

The internal space where learned representations live; operations here often correlate with semantics or generative factors.

A branch of ML using multi-layer neural networks to learn hierarchical representations, often excelling in vision, speech, and language.

When a model fits noise/idiosyncrasies of training data and performs poorly on unseen data.

When a model cannot capture underlying structure, performing poorly on both training and test data.

Fraction of correct predictions; can be misleading on imbalanced datasets.

Of true negatives, the fraction correctly identified.

Penalizes confident wrong predictions heavily; standard for classification and language modeling.

A parameterized function composed of interconnected units organized in layers with nonlinear activations.

A high-priority instruction layer setting overarching behavior constraints for a chat model.

A high-capacity language model trained on massive corpora, exhibiting broad generalization and emergent behaviors.

Retrieval based on embedding similarity rather than keyword overlap, capturing paraphrases and related concepts.

Constraining outputs to retrieved or provided sources, often with citation, to improve factual reliability.

Measure of consistency across labelers; low agreement indicates ambiguous tasks or poor guidelines.

Policies and practices for approving, monitoring, auditing, and documenting models in production.

How many requests or tokens can be processed per unit time; affects scalability and cost.

Hardware resources used for training/inference; constrained by memory bandwidth, FLOPs, and parallelism.

Tendency to trust automated suggestions even when incorrect; mitigated by UI design, training, and checks.

Mechanisms for retaining context across turns/sessions: scratchpads, vector memories, structured stores.

Error due to sensitivity to fluctuations in the training dataset.

Assigning labels per pixel (semantic) or per instance (instance segmentation) to map object boundaries.