Fetching from the wire…
01
02
03
04
05
06
07
08
Research2026-06-29 · source-backed
David Steinmann, Antonia Wüst, and Kristian Kersting's benchmark surfaces logical inconsistencies in interpretable models using "truly hard-negative" examples that earlier benchmarks missed. For anyone whose safety story depends on a model's chain-of-thought being trustworthy, this matters: it checks whether the reasoning is genuinely consistent or just superficially plausible.
Each link below shares sources, entities, or timing with this story.
Same source domain / Shared topic
Reported by the same outlet (arxiv.org); overlapping topics (actually, benchmark, model, reasoning).
Reported by the same outlet (arxiv.org); overlapping topics (actually, benchmark, model, whether).
Reported by the same outlet (arxiv.org); overlapping topics (actually, benchmark, check, whether).
Reported by the same outlet (arxiv.org); overlapping topics (anyone, benchmark, check, whether).
Reported by the same outlet (arxiv.org); overlapping topics (benchmark, chain-of-thought, model, reasoning).
Reported by the same outlet (arxiv.org); overlapping topics (chain-of-thought, earlier, model, reasoning).
Same source domain / Shared topic / Tension
Reported by the same outlet (arxiv.org); overlapping topics (benchmark, model); pushes against this story (competes).
Reported by the same outlet (arxiv.org); overlapping topics (model, reasoning); pushes against this story (against).