sequelbox/Raiden-Mini-DeepSeek-V3.2-Speciale

---
license: apache-2.0
tags:
- raiden
- raiden-mini
- creative
- analytical
- reasoning
- rational
- deepseek
- v3.2-speciale
- v3.2
- deepseek-v3.2-speciale
- 685b
language:
- en
task_categories:
- text-generation
size_categories:
- 1K<n<10K
configs:
- config_name: default
data_files: Raiden_Mini_DS3.2_Speciale.csv
- config_name: comparison
data_files: Raiden_Mini_Comparative.csv
---
**[Click here to support our open-source dataset and model releases!](https://huggingface.co/spaces/sequelbox/SupportOpenSource)**

**Raiden-Mini-DeepSeek-V3.2.Speciale** is a dataset containing creative-reasoning and analytic-reasoning responses, testing the limits of [DeepSeek-V3.2.Speciale's](https://huggingface.co/deepseek-ai/DeepSeek-V3.2-Speciale) reasoning skills!

This dataset contains:

- a default subset of ~8k 'creative_content' and 'analytical_reasoning' prompts from [sequelbox/Raiden-DeepSeek-R1](https://huggingface.co/datasets/sequelbox/Raiden-DeepSeek-R1), with all responses generated by [DeepSeek V3.2 Speciale.](https://huggingface.co/deepseek-ai/DeepSeek-V3.2-Speciale)
- provides an unfiltered look into the reasoning skills of DeepSeek V3.2 Speciale.
- suitable to complement other reasoning datasets in finetuning.
- a comparison subset with the V3.2 Speciale responses alongside responses from V3.2.
- allows for a direct comparison between the V3.2 Speciale and V3.2 models, using the same prompts.
- some V3.2 responses (~30) have failed automatic quality checks, as a warning for any finetuning use with the V3.2 column. (No V3.2 Speciale responses failed these checks.)

**Responses have not been edited at all:** the Raiden-Mini dataset strives to accurately represent V3.2 Speciale and V3.2. Potential issues may include inaccurate answers and infinite thought loops. Raiden is presented as-is to be used at your discretion.

Users should consider applying their own sub-filtering and manual examination of the dataset before use in training.

Do as you will.