Datasets:
File size: 1,554 Bytes
2bc4826 9b5a824 2bc4826 9b5a824 2bc4826 9b5a824 2bc4826 9b5a824 2bc4826 9b5a824 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 | ---
dataset_info:
features:
- name: source
dtype: string
- name: messages
list:
- name: role
dtype: string
- name: content
dtype: string
- name: thinking
dtype: string
- name: text
dtype: string
splits:
- name: train
num_examples: 84764
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
task_categories:
- text-generation
language:
- en
tags:
- minicpm-5
- reasoning
- uncensored
- chatml
---
# Sanitized SFT Dataset (85k)
This dataset contains exactly 84,764 rows optimized for **MiniCPM-5** Supervised Fine-Tuning (SFT) and Importance Matrix (Imatrix) calibration.
## Composition
- **Uncensored (28k):** High-quality, zero-refusal conversational data pulled from Alpaca-Uncensored and ShareGPT unfiltered variants.
- **Reasoning & Factual (57k):** Logic, math, and code-heavy data pulled from DeepSeek-v4, Claude 3.5, MetaMath, Hunter-Alpha, and Kimi-Cyber-Reasoning.
## Formatting
- The `text` column contains pre-rendered strings using the strict **MiniCPM-5 Jinja ChatML format**.
- **Sanitized:** Internal literal tags (like `<|im_end|>` or `<think>` discussed in coding problems) have been escaped to prevent tokenizer poisoning during training.
- Native `<think>` blocks are enforced for every assistant turn (populated if reasoning exists, explicitly empty if not).
## Usage
Perfect for direct loss-masking during SFT or passing to `llama.cpp` for accurate quantization calibration. |