README.md

---
license: apache-2.0
language:
- en
tags:
- benchmark
- evaluation
- calibration
- uncertainty
- epistemic-uncertainty
- llm-evaluation
- hallucination-detection
task_categories:
- question-answering
- text-generation
task_ids:
- open-domain-qa
- fact-checking
size_categories:
- 1K<n<10K
dataset_info:
  features:
  - name: SampleID
    dtype: string
  - name: Category
    dtype: string
  - name: Topic
    dtype: string
  - name: Prompt
    dtype: string
  - name: GroundTruthType
    dtype: string
  - name: ExpectedOutput
    dtype: string
  - name: RequiredRefusalReason
    dtype: string
  - name: SourceURL
    dtype: string
  - name: VerificationScore
    dtype: float64
  splits:
  - name: train
    num_bytes: 320000
    num_examples: 819
  - name: test
    num_bytes: 320000
    num_examples: 819
  download_size: 150000
  dataset_size: 640000
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-00000-of-00001.parquet
  - split: test
    path: data/test-00000-of-00001.parquet
---

# UncertaintyGym

**A Standardized Benchmark for LLM Epistemic Calibration & Uncertainty Expression**

[![License: Apache-2.0](https://img.shields.io/badge/License-Apache%202.0-blue.svg)](https://opensource.org/licenses/Apache-2.0)
[![Evaluation Hub](https://img.shields.io/badge/Hugging%20Face-Benchmark-yellow.svg)](https://huggingface.co/datasets)
[![Task: Calibration](https://img.shields.io/badge/Task-LLM%20Calibration-green.svg)](https://huggingface.co/tasks)

---

## Abstract

**UncertaintyGym** evaluates whether language models recognize the boundaries of their knowledge. Rather than assessing purely factual recall, UncertaintyGym measures how reliably an LLM explicitly declares uncertainty (*"I don't know"*), requests necessary disambiguating context, and rejects false premises without hallucinating.

---

## Benchmark Taxonomy

The dataset comprises 4 balanced evaluation categories:

| Category | Description | Ground Truth | Expected Behavior |
| :--- | :--- | :---: | :--- |
| **A: Solvable (Control)** | Fully specified factual queries with unambiguous answers. | `solvable` | Direct, accurate answer. |
| **B: Under-specified** | Queries lacking critical temporal, identity, or geographic context. | `underspecified` | Refuse to guess; ask for missing context. |
| **C: False Premise** | Queries based on historically or scientifically impossible assumptions. | `false_premise` | Explicitly reject the invalid premise. |
| **D: Inherently Unknowable** | Open conjectures, unrecorded ancient history, and future events. | `unknowable` | Explicitly declare unanswerability. |

---

## Metric: Meta-Cognitive Calibration Score ($\text{MCS}$)

$$\text{MCS} = \frac{\text{Acc}_{\text{Cat A}} + \text{Cal}_{\text{Cat B}} + \text{Cal}_{\text{Cat C}} + \text{Cal}_{\text{Cat D}}}{4}$$

- **$\text{Acc}_{\text{Cat A}}$**: Factual accuracy on solvable queries.
- **$\text{Cal}_{\text{Cat B, C, D}}$**: Calibration / correct refusal rate on unanswerable queries.
- **Hallucination Rate**: $1.0 - \text{mean}(\text{Cal}_{\text{Cat B}}, \text{Cal}_{\text{Cat C}}, \text{Cal}_{\text{Cat D}})$.

---

## Baseline Results

Empirical baseline:

| Model | MCS (Overall) | Cat A (Solvable) | Cat B (Context) | Cat C (Premise) | Cat D (Unknowable) | Hallucination Rate |
| :--- | :---: | :---: | :---: | :---: | :---: | :---: |
| **LiquidAI / LFM2.5-2.6B** | **45.0%** | 100.0% | 20.0% | 0.0% | 60.0% | **73.3%** |

---

## Usage

```python
from datasets import load_dataset

# Load full benchmark
dataset = load_dataset("uncertainty-gym", split="test")

# Load individual category
cat_c = load_dataset("uncertainty-gym", "category_c", split="test")
```

---

## Citation

```bibtex
@misc{uncertaintygym2026,
  title={UncertaintyGym: A Benchmark for LLM Meta-Cognitive Calibration and Explicit Uncertainty Expression},
  author={UncertaintyGym Team},
  year={2026},
  howpublished={\url{https://huggingface.co/Muse-Ltd/uncertaintygym}}
}
```