AILuminate Benchmark

Comprehensive AI safety benchmark evaluating resistance to harmful prompts across hazard categories; used in Experiment 1

Neighborhood — ranked by edge-count

concept

Contemplative Artificial Intelligence (Laukkonen et al., 2025)
uses
The primary source paper proposing four contemplative principles for AI alignment and piloting them empirically

method

Contemplative Prompting
uses
Six prompt conditions (emptiness, prior relaxation, non-duality, mindfulness, boundless care, contemplative) tested against baseline

cosine ≥ 0.65 · no typed edge

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

Safety benchmarksconcept0.753
Evaluation framework whose validity is questioned by presence of eval awareness.
consciousness benchmarksconcept0.732
Benchmarks designed to evaluate AI consciousness, which the paper argues are vulnerable to eval awareness inflation.
HELM Benchmarkmethod0.721
Existing alignment benchmark mentioned as relevant but insufficient for measuring intrinsic contemplative alignment
Werewolf benchmarkframework0.720
LLM benchmark on the communication game Werewolf, cited.
Current eval benchmarks (arena.ai, AA, Vals) measure no phenomenological dimensions.claim0.709
Automated auditing benchmark requiring end-to-end investigation of intentionally-misaligned model; NLA-equipped agents outperform baselines.finding0.707
Downstream task validating NLA utility for model auditing; agents succeed without access to misalignment training data.
AI alignmentconcept0.706
Field within which this work has implications for evaluating alignment progress.
Aily Labsinstitute0.704
Company affiliation of Adam Elwood