ANIMA: Animal Norms In Moral Assessment for AI Evaluation

ANIMA is a dataset of 26 prompts and 13 ethical dimensions for evaluating AI model moral reasoning about animal welfare. It was created by sentientfutures and introduced in the paper 'Alignment Midtraining for Animals' (Brazilek & Tidmarsh, 2026). The dataset was last updated on Hugging Face in May 2026.

Use Cases

• Benchmarking AI model performance on moral reasoning tasks based on the 13 ethical dimensions.
• Evaluating the consistency of an AI's ethical judgments across the 26 distinct animal welfare prompts.
• Training or fine-tuning models for improved ethical alignment on topics related to animal treatment.
• Conducting research on the intersection of AI alignment and animal ethics based on the defined moral assessment framework.

Strengths

• Contains 26 distinct prompts for evaluation.
• Defines 13 specific ethical dimensions for structured moral assessment.
• Associated with a formal academic paper, providing research context.

Limitations

• Row count and overall data scale are unknown.
• Column-level documentation is absent; field semantics must be inferred after download.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

sentientfutures on Hugging Face.

Collection Method

Likely created for the purpose of AI evaluation as described in the associated paper.

Freshness

Last updated 2026-05-02 17:40:01; freshness should be verified.