Publications

D. N. Yaldiz*, Y. Bakman*, S. Avestimehr, S. P. Karimireddy, “Hair-Trigger Alignment: Black-Box Evaluation Cannot Guarantee Post-Update Alignment”,  in submission 2026.

D. N. Yaldiz, E. Spiliopoulou, Z. Qi, S. Doss, S. Varia, N. Pappas, “Balancing Classification and Calibration Performance in Decision-Making LLMs via Calibration Aware Reinforcement Learning”,  ACL Findings 2026.

N. Wang, D. N. Yaldiz, Y. Bakman, S. P.  Karimireddy, “Conformal Prediction Adaptive to Unknown Subpopulation Shifts”,  in submission 2025.

Y. Bakman, S. Kang, Z. Huang, D. N. Yaldiz, C. G. Belém, C. Zhu, A. Kumar, A. Samuel, S. Avestimehr, D. Liu, S. P. Karimireddy, “Uncertainty as Feature Gaps: Epistemic Uncertainty Quantification of LLMs in Contextual Question-Answering”,  ICLR 2026.

A. Ziashahabi, Y. Bakman, D. N. Yaldiz, M. El-Khamy, S. P. Karimireddy, S. Avestimehr, “Reject Only Critical Tokens: Pivot-Aware Speculative Decoding”,  NeurIPS 2025 Efficient Reasoning Workshop.

D. N. Yaldiz*, Y. Bakman*, S. Kang, A. Ozis, H. Yildiz, M. Shah, Z. Huang, A. Kumar, A. Samuel, D.Liu, S. P. Karimireddy, S. Avestimehr, “TruthTorchLM: A Comprehensive Library for Predicting Truthfulness in LLM Outputs”,  EMNLP System demonstrations 2025.

D. N. Yaldiz*, Y. Bakman*, S. Kang, T. Zhang, B. Buyukates, S. Avestimehr, S. P.  Karimireddy, “Reconsidering LLM Uncertainty Estimation Methods in the Wild”,  ACL 2025.

M. Oğuz, Y. Bakman, D. N. Yaldiz, “Un-considering Contextual Information: Assessing LLMs’ Understanding of Indexical Elements”, ACL Findings 2025.

D. N. Yaldiz*, Y. Bakman*, B. Buyukates, C. Tao, A. Ramakrishna, D. Dimitriadis, S. Avestimehr, “Do Not Design, Learn: A Trainable Scoring Function for Uncertainty Estimation in Generative LLMs”, NAACL Findings 2025.

E. Mushtaq, D. N. Yaldiz, Y. Bakman, J. Ding, C. Tao, D. Dimitriadis, S. Avestimehr, “CroMo-Mixup: Augmenting Cross-Model Representations for Continual Self-Supervised Learning”, ECCV, 2024.

Y.Bakman, D. N. Yaldiz, B. Buyukates, C. Tao, D. Dimitriadis, S. Avestimehr, “MARS: Meaning-Aware Response Scoring for Uncertainty Estimation in Generative LLMs”, ACL, 2024.

D. N. Yaldiz*, Y.Bakman*, Y. Ezzeldin, S. Avestimehr, “Federated Orthogonal Training: Mitigating Global Catastrophic Forgetting in Continual Federated Learning”, ICLR, 2024.

D. N. Yaldiz, T.Zhang, S. Avestimehr, “Secure Federated Learning against Model Poisoning Attacks via Client Filtering”, ICLR Workshop on Backdoor Attacks and Defenses in Machine Learning, 2023.