Fallacy: Logical Fallacy Detection Dataset with 138k+ Examples

A dataset for detecting 14 types of logical fallacies in English text, created by kuwrom. It contains 138,574 rows for multi-class classification and 25,068 instruction examples for fine-tuning. The dataset was last updated on June 3, 2026.

Use Cases

• Multi-class text classification based on 14 fallacy types.
• Instruction fine-tuning (SFT) for chat models based on the 'messages' schema.
• Training models to identify argumentative flaws in written text.

Strengths

• Provides 138,574 labeled examples for classification tasks.
• Includes a separate configuration with 25,068 examples formatted for instruction fine-tuning.
• Covers 14 distinct types of logical fallacies.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count for the classification configuration is provided, but other metadata like file formats and license are unknown.

Provenance

Source

huggingface

Collection Method

Likely compiled from short, single-statement examples labelled by fallacy type.

Freshness

Last updated 2026-06-03 23:01:36; freshness should be verified.