Get started
Author

Cameron Tice

Observed name: Tice, Cameron
OpenAlex ID: A5123233154
Alignment Pretraining: AI Discourse Causes Self-Fulfilling (Mis)alignment
2026 · Tice, Cameron, Radmard, Puria, et al. · arXiv (Cornell University) · 38
PDF
Large language models can learn and generalize steganographic chain-of-thought under process supervision
2025 · Skaf, Joey, Ibanez-Lissen, Luis, et al. · arXiv (Cornell University) · 34
PDF
Noise Injection Reveals Hidden Capabilities of Sandbagging Language Models
2025 · Cameron Tice, Philipp Kreer, et al. · neural information processing systems · 19
PDF
Chain-of-thought obfuscation learned from output supervision can generalise to unseen tasks
2026 · Hadida, Nathaniel Mitrani, Bhanji, Sassan, et al. · arXiv (Cornell University) · 6
PDF
Constitutional Midtraining: Content Presence Drives Alignment Gains
2026 · Cho, Desiree, Tice, Cameron, et al. · arXiv (Cornell University) · 5
PDF
Inoculation Midtraining with Learned Neologisms
2026 · O'Brien, Kyle, Young, Edward James, et al. · arXiv (Cornell University) · 4
PDF
Diagnosing Pathological Chain-of-Thought in Reasoning Models
2026 · Liu, Manqing, Williams-King, David, et al. · arXiv (Cornell University) · 2
PDF
Large language models can learn and generalize steganographic chain-of-thought under process supervision
2025 · ROBERT MC CARTHY, Joey SKAF, et al. · neural information processing systems · 1
A transformer architecture alteration to incentivise externalised reasoning
2026 · Pavlova, Elizabeth, Koroliuk, Mariia, et al. · arXiv (Cornell University)
PDF
Enhancing Biosecurity with Language Models: Defining Research Directions
2024 · Michael Chen, Martin Holub, et al. · SSRN Electronic Journal
Works count: 10
Citation count: 108
H-index: 5
I10-index: 3
Year
Topic
Type