FAQs
FAQ: Why are my crosswalks not showing up in reports?
If your crosswalks are not showing up in the reports, it may be because not all controls (Safety, Alignment, Cost, and Carbon) were selected. Crosswalk results will not be generated if any of these controls are missing, as they will always be marked as failed. This is because the two currently available crosswalks, UK FCA and NIST AI RMF, contain statements that depend on passing all listed controls.
Recommendation: To see crosswalk results, please ensure that all controls are selected and then retry the scan. If the crosswalk results still do not generate, please contact our support team at help@trustwise.ai for further assistance.
FAQ: Why have my crosswalks failed?
Crosswalks are currently aggregate results of all Safety and Alignment metrics. All metrics need to pass evaluations for the crosswalk to be marked as passed. This often does not happen as at least one metric is often under the required threshold.
Recommendation: Check all individual Safety and Alignment metrics to identify which one(s) might be failing. Adjust your settings or rerun your scan to see if the issue resolves.
FAQ: Why does my error say I have a 5 scan limit?
If you are receiving an error message about having a 5 scan limit, you are likely on the Quickstart free tier, which provides a maximum of 5 scans.
Recommendation: If you need more scans, consider upgrading to a paid plan or contact our support team for more information.
FAQ: Why does it say I have reached the 5 scan limit when I only see 4 rows in Command Center?
Even if you only see 4 (or less) scans listed in the Command Center, it might be the case that you have ran 5 scans as failed scans are also counted towards the 5 scan limit. If a scan fails and you rerun it, this is considered as two separate scans.
Recommendation: If you want to understand why a scan failed, please check your document has text and not only images, that API key is valid, and that your LLM is not outdated. If you still face issues, contact our support team at help@trustwise.ai for assistance.
FAQ: LLMs are present in the LLM dropdown, but I cannot register them even though my API key is valid.
This may happen because the Optimize:AI scan requires tokenizers from Hugging Face. If the tokenizers for a selected model are not openly available, we cannot calculate token counts or costs, making those LLMs unavailable for scans. Additionally, some LLMs listed by providers may not yet be openly available for usage in scans.
Recommendation: Ensure that the LLM and tokenizer are openly available. If you still have issues, contact our support team.
FAQ: Why do scans fail with NVIDIA models when Cost is selected?
NVIDIA has not yet publicly disclosed the costs of their models, which means we are unable to calculate costs for scans that include the Cost control.
Recommendation: Please rerun your scan without selecting the "Cost" control when using NVIDIA models.
FAQ: Why are my reports or charts not generated after a scan failure?
If your scan fails due to a significant issue, such as the model being unavailable or hitting rate limits, the system might be unable to generate reports or charts. This is because the scan does not complete successfully, leaving insufficient data for report generation.
Recommendation: Check if the selected model is available and that you have not exceeded API rate limits. If the issue persists, ensure that your API key is valid and that your document meets the required quality standards.
For any issues that were not addressed, please reach out to our support team at help@trustwise.ai for assistance.