Evaluation & Testing

Agent & Prompt Frameworks

A library for building LLM applications that also auto-tunes them, optimizing prompt text and few-shot demonstrations against a training set. It aims to make prompt tuning as routine as model training.

Evaluation & Testing

A pytest-style evaluation framework for LLM output with ready-made metrics for hallucination, answer relevancy, faithfulness and task completion. Tests run locally and can be pushed to the Confident AI platform.

Agent & Prompt Frameworks

A Python framework that replaces hand-written prompt strings with declarative modules and signatures, then compiles them against your data. Optimizers search for the wording and few-shot examples that score best on a metric you define.

Evaluation & Testing

A vulnerability scanner for language models that probes a target with hundreds of attack prompts covering jailbreaks, prompt injection, data leakage and toxic output, then reports which probes succeeded.

Agent & Prompt Frameworks

An optimizer that improves prompts and other text artifacts through reflective evolution, reading execution traces and natural-language feedback to propose edits. It integrates with DSPy and other pipelines.

Agent & Prompt Frameworks

deepset's orchestration framework for retrieval and agent pipelines built from composable components. Prompt builders are pipeline nodes, so prompt text is versioned with the rest of the pipeline definition.

Evaluation & Testing

An evaluation framework from the UK AI Security Institute with built-in components for prompt engineering, tool use, multi-turn dialogue and model-graded scoring, plus a log viewer for inspecting runs.

Evaluation & Testing

A Microsoft research framework for evaluating language models with a focus on prompt robustness, including adversarial prompt attacks and prompt engineering method comparisons. The repository is archived and kept for reference.

Evaluation & Testing

A command-line and library test runner for prompts, models and RAG pipelines. Test cases are declared in YAML, run side by side across providers, and the project also performs red-team scans for prompt injection and jailbreaks.

Prompt Managers & Versioning

A prompt management platform where non-engineers can edit prompts in a visual registry while engineers pull versions at runtime. It also logs requests and runs evaluation pipelines against prompt versions.

Agent & Prompt Frameworks

A research library that treats natural-language critique as a gradient, backpropagating textual feedback through a pipeline to improve prompts and other text variables. It mirrors the autograd style familiar from PyTorch.

Evaluation & Testing

An instrumentation and evaluation library that wraps an application, records each call and scores it with feedback functions such as groundedness and relevance. Results are browsable in a local dashboard.

Evaluation & Testing

A library of reference-free metrics for retrieval-augmented pipelines, scoring faithfulness, context precision and answer relevance. It also generates synthetic test sets from your own documents.

Evaluation & Testing

A holistic evaluation framework from Stanford CRFM that scores models across many scenarios and metrics including accuracy, calibration, robustness and bias, and publishes public leaderboards.

Evaluation & Testing

An open-source framework for evaluating and monitoring AI systems, with more than a hundred built-in metrics plus LLM-as-judge evaluators and drift detection, rendered as interactive reports.

Evaluation & Testing

EleutherAI's harness for few-shot benchmarking of language models across hundreds of academic tasks. It standardises prompt formatting so results are comparable between models and backends.

Agent & Prompt Frameworks

An OpenAI-compatible inference proxy that applies reasoning techniques such as chain-of-thought variants, best-of-n sampling and self-consistency at request time to improve answer quality.

Evaluation & Testing

A modular evaluation library for LLM pipelines that combines deterministic metrics, semantic metrics and LLM-based judgement at each stage of a chain rather than only at the final answer.

Evaluation & Testing

A validation library that began in tabular machine learning and now covers LLM applications, running suites of automated checks over data and model output and producing shareable reports.

Agent & Prompt Frameworks

A prompt tuning framework built around intent-based prompt calibration. It generates challenging edge-case examples, collects judgements on them and rewrites the prompt to close the gap.

Evaluation & Testing

An automatic evaluator for instruction-following models that uses an LLM judge to compare outputs against reference responses, with published leaderboards and length-controlled scoring.

Evaluation & Testing

An open-source platform for scoring generative AI responses on dimensions like factual accuracy, retrieval quality and tone, with root-cause analysis to show which pipeline step failed.

Evaluation & Testing

An NVIDIA toolkit for adding programmable rails to conversational applications. Rails are written in a dedicated modelling language and control topics, dialogue flow and unsafe output.

Evaluation & Testing

OpenAI's framework and open registry for benchmarking models and prompt chains. Evals are declared as YAML plus data files, and the registry collects community-contributed benchmarks.

Evaluation & Testing

A Python package from CVS Health for uncertainty quantification in language models, providing hallucination detection scorers based on response consistency and token probabilities.

Prompt Managers & Versioning

A prompt management platform aimed at product teams, combining a playground for iterating on prompts with test suites that run those prompts against saved cases before deployment.

Agent & Prompt Frameworks

A Microsoft Research framework for end-to-end optimization of AI workflows, propagating execution feedback back through a computation graph to update prompts, code and parameters.

Prompt Managers & Versioning

Comet's open-source platform for tracing, evaluating and monitoring LLM applications, including a prompt library with versioning and an agent optimizer for improving prompt text.

Prompt Managers & Versioning

An open-source workspace where engineers and subject-matter experts iterate on prompts and agent configurations side by side, with versioning, evaluation runs and observability.

Agent & Prompt Frameworks

Google's open-source Python toolkit for building and deploying agents, with agent instructions, tools, sessions and a built-in evaluation runner for checking agent trajectories.

Evaluation & Testing

A desktop application and Python library for building AI systems, covering prompt iteration, dataset collection, evaluations, synthetic data and fine-tuning in one project file.

Prompt Managers & Versioning

Arize's open-source observability tool that runs locally or self-hosted, offering OpenTelemetry tracing, a prompt playground with saved prompt versions and built-in evaluators.

Evaluation & Testing

An open-source visual environment for stress-testing prompts. Prompt variants, models and evaluators are wired together as nodes, and results are compared in charts and tables.

Evaluation & Testing

An evaluation and observability platform for generative applications, with automated metrics for hallucination and instruction adherence plus guardrails for production traffic.

Evaluation & Testing

An open-source testing library that automatically scans LLM applications and agents for issues such as hallucination, prompt injection, harmful content and robustness failures.

Prompt Managers & Versioning

An open-source observability and gateway platform that logs every model request through a one-line proxy change, with prompt versioning, caching and experiment tooling on top.