aieng-lab/gradiend-mib-ioi

GRADIEND MIB IOI Data

Indirect-object-identification cloze data adapted from the Mechanistic Interpretability Benchmark (MIB), used in aieng-lab/iend-study.

Usage

from datasets import load_dataset

ds = load_dataset("aieng-lab/gradiend-mib-ioi", split="train")

Splits: train, validation, test.

Dataset Details

Dataset Description

Each example contains an indirect object and a subject name, e.g. As Carl and Maria left the consulate, Carl gave a fridge to [MASK]. The indirect object (Maria, class IO) is the factual target and the subject (Carl, class SUBJECT) the alternative, yielding the one-sided IO-versus-subject feature used in the study.

Dataset Structure

One config (default); 10,000 / 10,000 / 1,000 rows (train / validation / test).

• masked: the input text with a single [MASK] slot to be predicted
• label: the factual target token(s) for the slot
• label_class: class of the factual target
• alternative: the alternative (counterfactual) target for the slot
• alternative_class: class of the alternative target
• feature_class_id: identifier of the feature class (equal to label_class)
• split: train, validation, or test
• pair_id, template, cf_masked, source: provenance, the IOI template, and the counterfactual text with subject and indirect object swapped

Dataset Sources

• Repository: https://github.com/aieng-lab/iend-study
• Original Data: mib-bench/ioi (MIT)

Dataset Creation

Adapted from mib-bench/ioi (MIT), distributed with the Mechanistic Interpretability Benchmark (Mueller et al., 2025). We keep the original MIB data split and add the IO-versus-subject target/alternative columns.

Bias, Risks, and Limitations

Templated, name-based data; it measures one narrow circuit-level behaviour. Names inherit the biases of the original benchmark.

Citation

@misc{drechsel2026gradientsinterventionsactivationsdetection,
      title={Gradients for Interventions and Activations for Detection: Targeted Feature Learning in Language Models}, 
      author={Jonathan Drechsel and Steffen Herbold},
      year={2026},
      eprint={2609.32355},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2609.32355}, 
}