Latest Ragas Scores

20-question hand-labeled AWS Bedrock RAG evaluation.

faithfulness0.9762
answer_relevancy0.5706
context_precision0.6417
context_recall0.5333
answer_correctness0.5285

Question-Level Results

QuestionRAG answer excerptfaithfulnessanswer_relevancycontext_precisioncontext_recallanswer_correctness
What is Amazon Bedrock?Based on the provided context, **Amazon Bedrock** is a **fully managed service** that provides **secure, enterprise-grade access to high-performing foundation models** from leading AI companies, enabling you to **build a1.00000.93261.00000.00000.7364
Which foundation model providers are available on Amazon Bedrock?Based on the provided context, the following foundation model providers are available on Amazon Bedrock: 1. **Amazon** (Nova models and Titan models) 2. **Anthropic** (Claude models) 3. **AI21 Labs** 4. **Cohere** 5. **1.00000.99780.50001.00000.6632
Does Amazon Bedrock support streaming responses?Yes, Amazon Bedrock supports streaming responses. This is indicated by the mention of the `InvokeModelWithResponseStream` API operation and the `ConverseStream` operation, as well as the `responseStreamingSupported` fiel1.00001.00001.00001.00000.5730
How do you invoke a model on Amazon Bedrock?Based on the provided context, I don't know the specific steps to invoke a model on Amazon Bedrock. The context only mentions that you should "**Submit prompts and generate responses with model inference**" operations pr1.00000.00000.50000.00000.5360
What inference parameters can be configured when calling a Bedrock model?Based on the provided context, I don't know based on the provided context. The context only provides an **overview/introduction** about inference request parameters and response fields for foundation models in Amazon Be0.66670.00000.00000.00000.1576
What are Amazon Bedrock Knowledge Bases used for?Based on the provided context, Amazon Bedrock Knowledge Bases are used to: 1. **Improve AI responses** through Retrieval Augmented Generation (RAG), a technique that uses information from data sources to improve the **r1.00001.00000.50001.00000.5925
What are Amazon Bedrock Agents?Based on the provided context, Amazon Bedrock Agents is a capability that allows you to **build and configure autonomous agents** in your application. These agents help end-users complete actions based on organization da1.00000.90050.50001.00000.7309
What is the purpose of Guardrails for Amazon Bedrock?Based on the provided context, the purpose of Amazon Bedrock Guardrails is to provide **configurable safeguards** to help build safe generative AI applications. Specifically, it: - Offers **comprehensive safety and priv1.00000.96440.83330.00000.6280
Can you customize models on Amazon Bedrock?Yes, you can customize models on Amazon Bedrock. Model customization is the process of providing training data to a model in order to improve its performance for specific use cases. Amazon Bedrock allows you to customize1.00000.96401.00000.00000.6609
Can Amazon Bedrock run batch inference?Yes, Amazon Bedrock can run batch inference. Based on the provided context, Amazon Bedrock supports batch inference jobs, including the ability to: - **Create** a batch inference job - **Monitor** batch inference jobs -1.00000.94361.00000.00000.5161
Is Amazon Bedrock serverless?I don't know based on the provided context. The provided context describes Amazon Bedrock as a "fully managed service" but does not specifically mention whether it is serverless or not.1.00000.00001.00000.00000.3789
What AWS service is commonly used to store batch inference input and output data for Bedrock?I don't know based on the provided context.1.00000.00000.00000.00000.0185
What is a model ID in Amazon Bedrock?Based on the provided context, a model ID in Amazon Bedrock is an identifier used **with the model invocation API operations**. The context mentions "a table of models and their IDs to use with the model invocation API o1.00000.00000.50000.00000.4269
Can Bedrock Knowledge Bases retrieve context from external data for RAG?Yes, Amazon Bedrock Knowledge Bases can retrieve context from external data sources for RAG (Retrieval Augmented Generation). According to the context, when a query is made, **"a knowledge base searches your data to find1.00000.90571.00001.00000.6151
What does contextual grounding in Guardrails help detect?Based on the provided context, contextual grounding checks in Guardrails help **detect hallucinations in model responses** that are either: - **Not grounded** (factually inaccurate or add new information) in the source, 1.00000.93950.33331.00000.8593
Does Amazon Bedrock require JavaScript in the browser to call InvokeModel?Based on the provided context, the requirement for JavaScript mentioned is specifically for using the **Amazon Web Services Documentation** website, not for calling the InvokeModel API itself. The context states: > "Jav1.00000.00001.00001.00000.5854
Does Bedrock support GPT-4o?I don't know based on the provided context. (The context mentions a model ID "openai.gpt-oss-120b-1:0" but does not provide any specific information about GPT-4o support in Amazon Bedrock.)1.00000.00000.00001.00000.2239
Are Amazon Bedrock Agents the same thing as ChromaDB?No, Amazon Bedrock Agents and ChromaDB are not the same thing. Based on the provided context, Amazon Bedrock Agents is a feature within Amazon Bedrock that allows you to **build and configure autonomous agents** that orc0.85710.99290.33330.66670.5070
Can every Bedrock model use exactly the same request body format?Based on the provided context, **no**, not every Bedrock model uses exactly the same request body format. The context states: > "When you make inference calls to models with the model invocation API operations, you incl1.00000.87061.00001.00000.6020
Does Amazon Bedrock automatically make private company data public?Based on the provided context, there is no information suggesting that Amazon Bedrock automatically makes private company data public. In fact, the context indicates the opposite — it describes Amazon Bedrock as providin1.00000.00000.83331.00000.5574