framework
active
framework:lm-character-traits-frameworkLM Character Traits Framework
The paper's central contribution: a formal behaviourist framework for attributing character traits to LMs based on input-output behaviour.
Neighborhood — ranked by edge-count
Papers (1)
paper
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- Core empirical question motivating Section 5 on stationary and reflective traits.
- Empirical question answered through experiments on model size, fine-tuning, and prompting.
- Finding replicated across multiple experiments.
- A quality attributable to an LM such as truthfulness, toxicity, sycophancy, or helpfulness, defined in terms of behavioural tendencies.
- GPT-4.1-mini-based evaluation protocol that scores trait expression in model responses on a 0-100 scale
- Automated scoring of trait expression on 0-100 scale using G20B as a local judge model
- Contentious concept operationalised behaviourally as consistent outputs of P or similar content when directly asked about P.
- The conceptual scheme that life is a universal, objective, graded property of all space, detectable by human feeling.