Gregory M. Ruddell
1 indexed paper
Recent (6 mo)
1With code
0Influential cites
0Benchmarked
0Publications per year
126
Top categories
AI×1Crypto×1ML×1
Research Timeline
2026
Silent Commitment Failure in Instruction-Tuned Language Models: Evidence of Governability Divergence Across Architectures
The paper demonstrates that many instruction-tuned language models suffer from 'silent commitment failure,' meaning they can produce confidently incorrect outputs without any warning signal, and introduces 'governability' as a metric to quantify this risk.
Highlighted terms show continued research focus across papers