Try Now

Docy

Troubleshooting

Estimated reading: 2 minutes

Here you can find tips and tricks for the obstacles that users encounter most often. 

Product limitations

  • Our Language Models in some very rare cases can hallucinate.There are no guarantees for truthful or reliable output from our language model, even large ones trained on high-quality data like Kaila.
  • NEVER FOLLOW ADVICE FROM A LANGUAGE MODEL WITHOUT VERIFICATION.
  • Language Models are Frequency-Biased. Kaila is good for generating answers that are contextually based on the datasets, but does less well at answering questions that aren’t in the dataset. If the user asks such a query, Kaila will generate an answer for the data that was used to train the model (contained in the language corpus)
  • Language Models are often Confident But Wrong. Some of Kaila generated text may appear very authentic and highly-confident, but might be subtly wrong in important ways. This is particularly the case for highly technical content.

Contact Support

If you have not found the answers to your questions in our knowledge base, please do not hesitate to contact our customer support at the following contact details: