The dataset viewer is taking too long to fetch the data. Try to refresh this page.
Server-side error
Error code: ClientConnectionError
Subsets and Splits
SQL Console for OpenAssistant/oasst2
Reveals the structure of conversations by pairing prompts with their corresponding responses, highlighting the ranking of assistant responses for English-language messages.
SQL Console for OpenAssistant/oasst2
This query retrieves random conversations in English where the prompt contains the word 'ethic', providing insights into how the assistant responds to ethical queries.
Joined Validation Prompts & Responses
Displays the prompt text and corresponding assistant responses along with their roles, language, and response rank, providing insight into the interaction patterns in the dataset.
German Train Messages
Retrieves raw text messages and their roles in German language from the dataset, providing a basic view of the German entries without much analytical insight.
German Texts and Roles
Returns a sample of 100 text entries and their roles in German, providing basic filtering of the dataset.
Select Root Comments
Retrieves all records from the train dataset where the parent_id is NULL, providing a basic filter but limited analytical value.
Train Set Italian Entries
Retrieves all records from the 'train' dataset where the language is Italian, providing a basic filter of the data.
English Prompts Without Parents
Retrieves English prompts from the dataset that do not have a parent, providing a basic view of initial prompts in English.
Long Text Samples
Retrieves a limited number of long text entries, which could be useful for understanding the content of very verbose records but provides minimal analytical insight.
Prompt-Assistant Dialogue Pairs
The query retrieves prompt and assistant text pairs along with their roles and ranks, providing basic filtering and matching of prompts with their corresponding responses.
Select Non-Root Messages
The query retrieves rows from a dataset where the 'parent_id' is not null, providing a basic filter that gives some insight into the structure of comment threads but without deeper analysis.
Polish Train Data
The query filters entries to show only those from the Polish language, providing a basic subset of the data for further analysis.
Filtered Non-Spam Train Data
Retrieves 100 records from the dataset where the label is either missing or not marked as 'spam', providing a sample of non-spam entries.