Snapshot 217915
Normalized text
Scripts and page chrome removed; this is what change detection compares.
The Eval Manifesto Encyclopedia Evalica A Absolute scoring Active observability Adversarial examples Agent Agent observability Alert / threshold Alignment Annotation Annotation schema Attachment B Baseline Baseline experiment Benchmark Brainstore C Caching Calibration CI/CD integration Coherence Confidence interval Coverage D Dashboard Data flywheel Dataset Dataset record Deep search Drift E Eval Edge case Error rate Eval harness Eval leakage Expected output (ground truth) Experiment F Factuality Failure mode Faithfulness Fallback Feedback loop Feedback signal flush() Function (deployed) G Gateway Golden dataset Groundedness H Hallucination Human review I Input Instrumentation Inter-annotator agreement (IAA) L Label LLM-as-a-judge Logs Loop M Metadata Model comparison Model drift Multimodal Multimodal dataset N Non-determinism O Offline evaluation Online evaluation (production scoring) P P50 / P95 / P99 (Percentiles) Pairwise evaluation Pass@k Playground Prompt Prompt (deployed) Q Quality gate R RAG (retrieval-augmented generation) RAG evaluation Reference-based scoring Reference-free scoring Regression testing Release criteria Remote evaluation Rubric S Safety Sampling rate Score distribution Scorer Semantic failure Service Level Indicator (SLI) Service Level Objective (SLO) Signal-to-noise ratio Slug Span Streaming T The Eval Manifesto Task (eval task) Time-to-first-token (TTFT) Token Token usage / cost tracking Tool call Topics Toxicity score Trace Trace Lake U User feedback W Wrap (provider wrap) ABCDEFGHIJKLMNOPQRSTUVWXYZ Encyclopedia Evalica A Absolute scoring Evaluation Active observability Observability Adversarial examples Datasets Agent Evaluation Agent observability Observability Alert / threshold Observability Alignment Evaluation Annotation Annotation Annotation schema Evaluation Attachment Tracing and instrumentation B Baseline Evaluation Baseline experiment Evaluation Benchmark Evaluation Brainstore Deployment C Caching Deployment Calibration Evaluation CI/CD integration Evaluation Coherence Evaluation Confidence interval Evaluation Coverage Datasets D Dashboard Observability Data flywheel Observability Dataset Datasets Dataset record Datasets Deep search Observability Drift Observability E Edge case Datasets Error rate Observability Eval Evaluation Eval harness Evaluation Eval leakage Evaluation Expected output (ground truth) Datasets Experiment Evaluation F Factuality Evaluation Failure mode Evaluation Faithfulness Evaluation Fallback Deployment Feedback loop Observability Feedback signal Evaluation flush() Datasets Function (deployed) Deployment G Gateway Deployment Golden dataset Datasets Groundedness Evaluation H Hallucination Evaluation Human review Annotation I Input Datasets Instrumentation Tracing and instrumentation Inter-annotator agreement (IAA) Evaluation L Label Annotation LLM-as-a-judge Evaluation Logs Observability Loop Evaluation M Metadata Datasets Model comparison Evaluation Model drift Observability Multimodal Evaluation Multimodal dataset Datasets N Non-determinism Evaluation O Offline evaluation Evaluation Online evaluation (production scoring) Observability P P50 / P95 / P99 (Percentiles) Observability Pairwise evaluation Evaluation Pass@k Evaluation Playground Evaluation Prompt Deployment Prompt (deployed) Deployment Q Quality gate Evaluation R RAG (retrieval-augmented generation) Evaluation RAG evaluation Evaluation Reference-based scoring Evaluation Reference-free scoring Evaluation Regression testing Evaluation Release criteria Evaluation Remote evaluation Evaluation Rubric Evaluation S Safety Evaluation Sampling rate Observability Score distribution Evaluation Scorer Evaluation Semantic failure Evaluation Service Level Indicator (SLI) Observability Service Level Objective (SLO) Observability Signal-to-noise ratio Evaluation Slug Deployment Span Tracing and instrumentation Streaming Deployment T Task (eval task) Evaluation Time-to-first-token (TTFT) Observability Token Deployment Token usage / cost tracking Observability Tool call Tracing and instrumentation Topics Observability Toxicity score Evaluation Trace Tracing and instrumentation Trace Lake Observability U User feedback Tracing and instrumentation W Wrap (provider wrap) Tracing and instrumentation Get started with Evals Braintrust is the observability platform for agents in production. By actively applying intelligence to agent traces and automatically surfacing critical patterns, Braintrust helps teams at Notion, Stripe, Box, OpenAI, and Cloudflare ship quality agents at scale. Start building