Unverified50% confidenceFactExact time
This counterfactual audit covers seven open-weight models (8B to 70B parameter scale) and four benchmarks spanning judicial reasoning and contract reasoning.
1
Sources
50%
Confidence
Long-term
Relevance
10/13/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe study proposes a leading-indicator model, nine design propositions, and seven falsifiable hypotheses, translating them into a testable research agenda.76% similarUnverifiedThe study uses counterfactual auditing to test the chain-of-thought faithfulness of legal AI: while keeping the case facts fixed and unchanged, it replaces the legal basis cited by the model and decodes the decision tendency from the model's hidden states, in order to judge whether the decision truly depends on the cited basis.69% similarUnverifiedThe team wrote automated acceptance scripts covering 63 test cases—including key player queries, factual trap questions, sensitive topic interception, and multi-turn conversations—all of which passed, with one-click regression testing supported.68% similarUnverifiedThe research team collected 103,939 scored model responses across ten configurations: eight LLMs with reasoning disabled plus two of those with maximum reasoning capability enabled.68% similarUnverifiedThe research team selected 8 models spanning different capability tiers and conducted cross-evaluations across 54 benchmarks66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1013275API
curl https://kongchang.com/api/v1/knowledge/claims/1013275MCP
get_claim(id=1013275)