Issue regarding active learning and questions with prompts(Custom question and answering [CQA])

[charith-phamax]

Github issues for C# /JS / Java/ Python should be used for bugs and feature requests. Use Stack Overflow for general "how-to" questions.

Sample information

1. Sample type: 48.customQABot-all-features
2. Sample language: nodejs
3. Sample name: Custom Q&A Bot with all features

Describe the bug

When using active learning, if a question is asked which is unclear, we get options "Did you mean? [question1][question2]" etc. If the selected question contains prompts, then on selecting anyone of those prompts, we are encountering an issue where the bot is not correctly identifying user intents and providing relevant responses. If the question selected from active learning card("did you mean" card), is asked directly and not from the card, then response is relevant.

To Reproduce

Steps to reproduce the behavior:

1. Clone the 48.customQABot-all-features sample
2. Keep active learning feature as TRUE for your "custom question and answering" project
3. Test the bot by asking questions which are unclear.
4. On getting "Did you mean? [question1][question2]" choose question which has prompts
5. Then select one of the prompt and evaluate the result.
6. To get correct response, directly ask the question without using the active learning card. then select the same prompt and evaluate the response.

Expected behavior

We expect the custom Q&A bot to correctly identify user questions and provide relevant responses based on the active learning and prompt features.

Screenshots

1. This screenshot contains active learning card.
   
2. This screenshot contains the prompts for the selected question.
   
3. This screenshot contains response to the selected prompt.(not relevant response)
   
4. This screenshot contains the question (asked directly).
   
5. This screenshot contains the response to the prompt selected.
   

Additional context

No additional context

[Amrit10737726]

It seems there's uncertainty about the relevance of tsv and how it triggers certain actions within the bot. plan is to reach out to the customer for clarification on their tsv and aims to revisit the original issue once more information is available. The complexity of the problem is acknowledged, but efforts are being made to address it effectively.