From 61cfb2a754ff2b667cd2e1fe89989885a2323be8 Mon Sep 17 00:00:00 2001 From: mshobana <51772793+mshobana@users.noreply.github.com> Date: Wed, 16 Apr 2025 14:02:55 +0530 Subject: [PATCH] Correct the description of Context relevance metric The description of Context Relevance metric was not right and it was not giving the explanation which it actual does. Updating it to the right one. --- docs/concepts/metrics/available_metrics/nvidia_metrics.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/concepts/metrics/available_metrics/nvidia_metrics.md b/docs/concepts/metrics/available_metrics/nvidia_metrics.md index 89a40704b..d39593461 100644 --- a/docs/concepts/metrics/available_metrics/nvidia_metrics.md +++ b/docs/concepts/metrics/available_metrics/nvidia_metrics.md @@ -76,7 +76,7 @@ Thus, the final **Answer Accuracy** score is **1**. ## Context Relevance -**Context Relevance** evaluates whether the **retrieved_contexts** (chunks or passages) are pertinent to the **user_input**. This is done via two independent "LLM-as-a-judge" prompt calls that each rate the relevance on a scale of **0, 1, or 2**. The ratings are then converted to a [0,1] scale and averaged to produce the final score. Higher scores indicate that the contexts are more closely aligned with the user's query. +**Context Relevance** evaluates how much the relevant **retrieved_contexts** (chunks or passages) are pertinent to the **user_input** (Completely non-relevant chunks are ignored from the score and won't affect the score. Only the chunks which are either completely relevant or partailly relevant are considered for the score calculation). This is done via two independent "LLM-as-a-judge" prompt calls that each rate the relevance on a scale of **0, 1, or 2**. The ratings are then converted to a [0,1] scale and averaged to produce the final score. Higher scores indicate that the contexts are more closely aligned with the user's query. - **0** → The retrieved contexts are not relevant to the user’s query at all. - **1** → The contexts are partially relevant.