9 terms
Showing all terms starting with R
A technique that enhances LLM outputs by fetching relevant external documents at query time, grounding responses in current or proprietary data.
An LLM variant that performs extended internal reasoning (chain-of-thought) before producing a final answer, improving accuracy on complex problems.
A machine learning paradigm where an agent learns to make decisions by receiving rewards or penalties based on actions taken in an environment.
An ensemble ML method that builds many decision trees on random data subsets and combines their predictions for better accuracy and robustness.
A neural network architecture with loops that allow information to persist across sequence steps, used for time series and language tasks before transformers.
A second-stage retrieval step that reorders an initial set of retrieved documents by relevance score before passing them to the LLM as context.
A framework of principles and practices ensuring AI systems are fair, transparent, accountable, private, and safe across their entire lifecycle.
Recall-Oriented Understudy for Gisting Evaluation - metrics comparing AI-generated text to reference summaries by measuring n-gram overlap.
Optimisations applied at inference time - such as caching, batching, and quantisation - to reduce the cost and latency of running AI models.
