docs: Add sentiment classification to LLM Analytics docs (#15195)

* docs(llm-analytics): add sentiment classification section to traces page

Documents the new on-demand sentiment classification feature from
PostHog/posthog#48277. Sentiment classifies user messages in traces
as negative/neutral/positive using a local ONNX model.

* docs(llm-analytics): add sentiment classification section to generations page

---------

Co-authored-by: inkeep[bot] <257615677+inkeep[bot]@users.noreply.github.com>

Author: inkeep[bot]

contents/docs/llm-analytics/generations.mdx

@@ -101,6 +101,10 @@ Spans are individual operations within a trace. Some spans represent generations
 
 You can automatically assess the quality of your generations using [evaluations](/docs/llm-analytics/evaluations). Evaluations use an LLM-as-a-judge approach to score outputs based on criteria like relevance, helpfulness, or safety.
 
+## Sentiment classification
+
+PostHog can classify the sentiment of user messages in your generations as negative, neutral, or positive. Sentiment is computed on-demand using a local model when you view a [trace](/docs/llm-analytics/traces) — no data is sent to third-party services.
+
 ## Event properties
 
 <GenerationEventSnippet />

contents/docs/llm-analytics/traces.mdx

@@ -27,6 +27,10 @@ Clicking on a trace opens a timeline of the interaction with all the generation
 <Caption>A trace presents LLM event data in a timeline, tree-structured view </Caption>
 
 
+## Sentiment classification
+
+PostHog can classify the sentiment of user messages in a trace as negative, neutral, or positive. Sentiment is computed on-demand using a local model when you view a trace – no data is sent to third-party services. Each trace gets an overall sentiment label and score, with a per-generation and per-message breakdown.
+
 ## AI event hierarchy
 
 ```mermaid