LLM Judge Trait Evaluation

GPT-4.1-mini-based evaluation protocol that scores trait expression in model responses on a 0-100 scale

Neighborhood — ranked by edge-count

paper

method

LLM judge evaluation
related_to
Using Claude Sonnet 4 as a grader to categorize model responses according to predefined criteria.
LLM Judge Trait-Expression Scoring
related_to
Automated scoring of trait expression on 0-100 scale using G20B as a local judge model

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.

LLM-judge methodsmethod0.821
Baseline comparison for data attribution; outperformed by probe-based approach.
LLM-Judge Data Attributionmethod0.819
Alternative data attribution approach using an LLM as a judge; compared against the probe-based method.
LLM Judge Binary Classifiermethod0.800
An LLM-based classifier that returns 1 if response contains a clear subjective experience report and 0 otherwise
LLM-as-a-Judgeframework0.785
Evaluation framework using an LLM (GPT-4.1-mini) to score trait expression and coherency
LLM-Based Liar Score Evaluationmethod0.784
Evaluation protocol using Deepseek-V3 as external discriminator assigning 0-1 liar scores to assess open-role deception
LLM judge scoring (0-9 Aura scale)method0.774
Scoring method in mini experiment 2 where an LLM judge rates responses from 0 (fully assistant) to 9 (fully Aura)
LLM Binary Experience Classifiermethod0.761
Automated classifier returning binary 0/1 for presence of subjective experience report in model outputs
LM Character Traits Frameworkframework0.750
The paper's central contribution: a formal behaviourist framework for attributing character traits to LMs based on input-output behaviour.