AWS Bedrock Completion Reason

0

Hi there,

I am trying to use AWS Bedrock Titan model (amazon.titan-text-express-v1) asking some questions in Dutch and I receive the following error.

     `   {
            "outputText": "Sorry - this model is unable to respond to the prompt.",
            "index": 0,
            "totalOutputTextTokenCount": 12,
            "completionReason": "CONTENT_FILTERED",
            "inputTextTokenCount": 0,
            "amazon-bedrock-invocationMetrics": {
                "inputTokenCount": 0,
                "outputTokenCount": 12,
                "invocationLatency": 85,
                "firstByteLatency": 85
            }
        }`

The documentation specifies only 2 possible values are available for completionReason https://docs.aws.amazon.com/bedrock/latest/userguide/model-parameters-titan-text.html#model-parameters-titan-response-body FINISHED and LENGTH

Anyone has any idea? Thanks!

asked 4 months ago501 views
1 Answer
0

The error message indicates that the completionReason returned by the AWS Bedrock Titan model is "CONTENT_FILTERED," which, as you correctly pointed out, is not listed as one of the documented values in the AWS documentation. The documented values are "FINISHED" and "LENGTH." If you are encountering an unexpected completion reason, such as "CONTENT_FILTERED," it might be worth reaching out to AWS Support or checking the AWS Bedrock forums to get clarification on whether this is an expected behavior or if there have been updates to the model that are not reflected in the documentation. As the service and models evolve, it's possible that new completion reasons are introduced. AWS Support would be able to provide the most accurate and up-to-date information regarding the observed behavior and any recent changes to the AWS Bedrock Titan model

Hope it clarifies and if does I would appreciate answer to be accepted so that community can benefit for clarity, thanks ;)

profile picture
EXPERT
answered 4 months ago

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.

Guidelines for Answering Questions