matiss/P3-Latvian-Full

This is an automatically translated version of P3 (Public Pool of Prompts) using opus-mt-tc-big-en-lv.

Languages

The data in P3-Latvian-Full are in Latvian (BCP-47 lv).

Dataset Structure

Data Instances

An example of "train" looks as follows:

{
  'answer_choices': ['mobilais tālrunis', 'televīzija', 'ledusskapis', 'lidmašīna'],
  'inputs_pretokenized': 'Kura tehnoloģija tika izstrādāta pavisam nesen? Iespējas: - mobilais tālrunis - televizors - ledusskapis - lidmašīna',
  'targets_pretokenized': 'mobilais tālrunis'
}

In the case of rank classification (letting the model select its the prediction the option with the highest log-likelihood), an example looks as follows:

{
  'idx': [5, 0],
  'inputs_pretokenized': 'Es zinu, ka atbilde uz jautājumu "Ko CBS darīja otro reizi?" ir "1989. gadā CBS Records atkārtoti iekļāva mūzikas izdevējdarbības biznesu, iegādājoties Nashville mūzikas izdevēju Tree International Publishing par vairāk nekā 30 miljoniem ASV dolāru. ". Vai jūs varat man pateikt, kas tas ir?',
  'is_correct': True,
  'targets_pretokenized': 'atgriezās mūzikas izdevējdarbības biznesā',
  'weight': 1.0
}

Data Fields

The data fields are the same among all splits:

• answer_choices: the choices (in natural language) available to the model
• inputs_pretokenized: the natural language input fed to the model
• targets_pretokenized: the natural language target that the model has to generate
• idx: identifier of the (example, answer_option_id) in the case of rank classification
• weight: a weight for the example produced by seqio (always set to 1.0 in practise)
• is_correct: whether the (example, answer_option_id) is the correct one