project:

type: meridianlabs-ai/inspect-docs

resources:

- CNAME

- listing/

inspect-docs:

title: Inspect

description: Open-source framework for large language model evaluations

image: /images/inspect.png

logo: images/aisi-logo.svg

favicon: favicon.svg

url: https://inspect.aisi.org.uk

module: inspect_ai

repo: UKGovernmentBEIS/inspect_ai

org: UK AI Security Institute

org_url: https://aisi.gov.uk

twitter: AISecurityInst

sidebar:

collapse-level: 1

navbar:

left:

- text: "User Guide"

menu:

- text: Basics

href: index.qmd

- text: Components

href: tasks.qmd

- text: Models

href: models.qmd

- text: Scoring

href: scoring.qmd

- text: Agents

href: agents.qmd

- text: Tools

href: tools.qmd

- text: Running

href: running.qmd

- text: Analysis

href: analysis.qmd

- text: Extensions

href: extensions.qmd

- text: "Reference"

href: reference/index.qmd

- text: "Extensions"

href: extensions/index.qmd

- text: "Evals"

href: evals/index.qmd

navigation:

- text: Basics

contents:

- text: Welcome

href: index.qmd

- href: tutorial.qmd

- href: options.qmd

- href: log-viewer.qmd

- text: VS Code

href: vscode.qmd

- text: Components

contents:

- href: tasks.qmd

- href: datasets.qmd

- href: solvers.qmd

- href: scorers.qmd

- text: Models

contents:

- href: models.qmd

- text: Providers

href: providers.qmd

- href: caching.qmd

- text: Concurrency

href: models-concurrency.qmd

- href: compaction.qmd

- href: fallbacks.qmd

- href: multimodal.qmd

- href: reasoning.qmd

- href: structured.qmd

- href: models-batch.qmd

- text: Scoring

href: scoring.qmd

contents:

- href: standard-scorers.qmd

- href: custom-scorers.qmd

- text: Model Grading

href: model-graded.qmd

- text: Scoring Metrics

href: metrics.qmd

- href: multiple-scorers.qmd

- href: scoring-workflow.qmd

- href: perplexity.qmd

- text: Agents

contents:

- href: agents.qmd

- href: react-agent.qmd

- href: deepagent.qmd

- text: Checkpointing

href: checkpointing.qmd

- text: Intervention

href: intervention.qmd

- href: multi-agent.qmd

- href: agent-custom.qmd

- href: agent-bridge.qmd

- href: human-agent.qmd

- text: Tools

contents:

- href: tools.qmd

- href: tools-standard.qmd

- text: MCP Tools

href: tools-mcp.qmd

- href: tools-custom.qmd

- href: sandboxing.qmd

- href: approval.qmd

- text: Running

href: running.qmd

contents:

- href: eval-sets.qmd

- href: parallelism.qmd

- href: handling-errors.qmd

- href: setting-limits.qmd

- href: control-channel.qmd

- href: early-stopping.qmd

- href: task-source.qmd

- href: tracing.qmd

- text: Analysis

href: analysis.qmd

contents:

- href: eval-logs.qmd

- text: Dataframes

href: dataframe.qmd

- href: scanners.qmd

- href: inspect-viz.qmd

- href: task-views.qmd

- text: Extensions

href: extensions.qmd

contents:

- text: Model APIs

href: extensions-model-api.qmd

- text: Components

href: extensions-components.qmd

- text: Sandboxes

href: extensions-sandboxes.qmd

- text: Approvers

href: extensions-approvers.qmd

- text: Hooks

href: extensions-hooks.qmd

- text: Filesystems

href: extensions-filesystems.qmd

format:

inspect-docs-html:

code-annotations: select

css: styles.css

execute:

enabled: false