Skip to content
TheChatArchiveThe Chat Archive
Human-written demonstration (OpenAssistant volunteers)· OpenAssistant (human-written)Hugging Face OpenAssistant Conversations (oasst2)

Can you explain, in details how you work, and include scientific reference ?

In OpenAssistant data, “assistant” turns were written by human volunteers as demonstrations. They are not model outputs.
  1. User

    Can you explain, in details how you work, and include scientific reference ?

  2. Assistant

    Sure! Open Assistant is an Open-Source project meant to give everyone access to a great chat based large language model. A language model (LM) is an AI model that is trained to predict the probability of a sequence of words in a given language. A language model can be trained in two ways: either as a Generative model, which can generate new text, or as a Discriminative model, which can rank the likelihood of a given text being correct.

    A large language model (LLM), uses a deep neural network with a Transformer architecture. The model is trained on a massive corpus of text data, where it learns to predict the next word in a sequence given the previous words. The LLM is then fine-tuned for specific NLP tasks, such as text classification, question answering, and machine translation, by training it on task-specific data. (Scientific Reference: https://research.google/pubs/pub46201/)

    LLMs have proven to be highly effective in various NLP tasks due to their ability to capture and leverage large-scale patterns in language, often outperforming traditional rule-based systems and other machine learning models.

    *Note - You can learn more details about the Open Assistant project here: https://docs.google.com/presentation/d/1iaX_nxasVWlvPiSNs0cllR9L_1neZq0RJxd6MFEalUY

Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.