autolabel
autolabel copied to clipboard
refuel-ai
Label, clean and enrich text datasets with LLMs.
Given that [OpenAI have added logprobs](https://platform.openai.com/docs/api-reference/chat/create#chat-create-logprobs) to chat completions, add that as an option for computing confidence, especially since text-davinci-003 is deprecated by OpenAI
**Is your feature request related to a problem? Please describe.** Currently, the autolabel function performs a strict comparison between the output generated by the language model and a predefined list...
Add the option to use GPT4 Turbo. 3x cheaper compared to GPT4. Adding gpt-4-turbo-preview or specifically gpt-4-0125-preview would be great Even though gpt-4-1106-preview seems part of the source code, it...
On running the example notebooks given in the repository, running into following issue: **To Reproduce** Simply run agent.plan() in any of the example notebooks **Expected behavior** Expected the code to...
# Pull Review Summary ## Description Closes #672 partially. Adds replicate as an LLM provider. Features implemented - - Enables inference of any model hosted on replicate - Calculates pricing...
**Is your feature request related to a problem? Please describe.** the problem arises when attempting to utilize the GPT-3.5-turbo-1106(or later) model within the current version of our application, stating that...
**Describe the bug** Change confidence tokenizer to LlamaTokenizer from Flan-T5 Tokenizer
For multilabel, extraction tasks
# Pull Review Summary ## Description Adding support for Wolfram Alpha LLM API: https://products.wolframalpha.com/llm-api/documentation The API allows to send an input query and parses the wolfram alpha output to be...
