Ali Modarressi

34 papers A* 2A 1Journal 20Unranked 11
YearRankTypeTitle / Venue / Authors
2026 — conf
Persistent Personas? Role-Playing, Instruction Following, and Safety in Extended Interactions.
EACL (Volume 1: Long Papers)
Pedro Henrique Luz de Araujo, Michael A. Hedderich, Ali Modarressi, Hinrich Schütze, Benjamin Roth
2026 J jnl
With Argus Eyes: Assessing Retrieval Gaps via Uncertainty Scoring to Detect and Remedy Retrieval Blind Spots.
CoRR
Zeinab Sadat Taghavi, Ali Modarressi, Hinrich Schütze, Andreas Marfurt
2025 — conf
Collapse of Dense Retrievers: Short, Early, and Literal Biases Outranking Factual Evidence.
ACL (1)
Mohsen Fayyaz, Ali Modarressi, Hinrich Schütze, Nanyun Peng
2025 J jnl
Collapse of Dense Retrievers: Short, Early, and Literal Biases Outranking Factual Evidence.
CoRR
Mohsen Fayyaz, Ali Modarressi, Hinrich Schütze, Nanyun Peng
2025 — conf
Do We Know What LLMs Don't Know? A Study of Consistency in Knowledge Probing.
EMNLP (Findings)
Raoyuan Zhao, Abdullatif Köksal, Ali Modarressi, Michael A. Hedderich, Hinrich Schütze
2025 J jnl
Do We Know What LLMs Don't Know? A Study of Consistency in Knowledge Probing.
CoRR
Raoyuan Zhao, Abdullatif Köksal, Ali Modarressi, Michael A. Hedderich, Hinrich Schütze
2025 A* conf
ImpliRet: Benchmarking the Implicit Fact Retrieval Challenge.
EMNLP
Zeinab Sadat Taghavi, Ali Modarressi, Yunpu Ma, Hinrich Schütze
2025 J jnl
ImpliRet: Benchmarking the Implicit Fact Retrieval Challenge.
CoRR
Zeinab Sadat Taghavi, Ali Modarressi, Yunpu Ma, Hinrich Schütze
2025 — conf
MEXA: Multilingual Evaluation of English-Centric LLMs via Cross-Lingual Alignment.
ACL (Findings)
Amir Hossein Kargaran, Ali Modarressi, Nafiseh Nikeghbal, Jana Diesner, François Yvon, Hinrich Schütze
2025 J jnl
MemLLM: Finetuning LLMs to Use Explicit Read-Write Memory.
Trans. Mach. Learn. Res.
Ali Modarressi, Abdullatif Köksal, Ayyoob Imani, Mohsen Fayyaz, Hinrich Schütze
2025 A* conf
NoLiMa: Long-Context Evaluation Beyond Literal Matching.
ICML
Ali Modarressi, Hanieh Deilamsalehy, Franck Dernoncourt, Trung Bui, Ryan A. Rossi, Seunghyun Yoon, Hinrich Schütze
2025 J jnl
NoLiMa: Long-Context Evaluation Beyond Literal Matching.
CoRR
Ali Modarressi, Hanieh Deilamsalehy, Franck Dernoncourt, Trung Bui, Ryan A. Rossi, Seunghyun Yoon, Hinrich Schütze
2025 J jnl
Persistent Personas? Role-Playing, Instruction Following, and Safety in Extended Interactions.
CoRR
Pedro Henrique Luz de Araujo, Michael A. Hedderich, Ali Modarressi, Hinrich Schütze, Benjamin Roth
2025 J jnl
Steering MoE LLMs via Expert (De)Activation.
CoRR
Mohsen Fayyaz, Ali Modarressi, Hanieh Deilamsalehy, Franck Dernoncourt, Ryan A. Rossi, Trung Bui, Hinrich Schütze, Nanyun Peng
2025 — conf
Time Course MechInterp: Analyzing the Evolution of Components and Knowledge in Large Language Models.
ACL (Findings)
Ahmad Dawar Hakimi, Ali Modarressi, Philipp Wicke, Hinrich Schütze
2025 J jnl
Time Course MechInterp: Analyzing the Evolution of Components and Knowledge in Large Language Models.
CoRR
Ahmad Dawar Hakimi, Ali Modarressi, Philipp Wicke, Hinrich Schütze
2024 J jnl
Consistent Document-Level Relation Extraction via Counterfactuals.
CoRR
Ali Modarressi, Abdullatif Köksal, Hinrich Schütze
2024 — conf
Consistent Document-level Relation Extraction via Counterfactuals.
EMNLP (Findings)
Ali Modarressi, Abdullatif Köksal, Hinrich Schütze
2024 J jnl
MEXA: Multilingual Evaluation of English-Centric LLMs via Cross-Lingual Alignment.
CoRR
Amir Hossein Kargaran, Ali Modarressi, Nafiseh Nikeghbal, Jana Diesner, François Yvon, Hinrich Schütze
2024 J jnl
MemLLM: Finetuning LLMs to Use An Explicit Read-Write Memory.
CoRR
Ali Modarressi, Abdullatif Köksal, Ayyoob Imani, Mohsen Fayyaz, Hinrich Schütze
2023 — conf
DecompX: Explaining Transformers Decisions by Propagating Token Decomposition.
ACL (1)
Ali Modarressi, Mohsen Fayyaz, Ehsan Aghazadeh, Yadollah Yaghoobzadeh, Mohammad Taher Pilehvar
2023 J jnl
DecompX: Explaining Transformers Decisions by Propagating Token Decomposition.
CoRR
Ali Modarressi, Mohsen Fayyaz, Ehsan Aghazadeh, Yadollah Yaghoobzadeh, Mohammad Taher Pilehvar
2023 A conf
Guide the Learner: Controlling Product of Experts Debiasing Method Based on Token Attribution Similarities.
EACL
Ali Modarressi, Hossein Amirkhani, Mohammad Taher Pilehvar
2023 J jnl
Guide the Learner: Controlling Product of Experts Debiasing Method Based on Token Attribution Similarities.
CoRR
Ali Modarressi, Hossein Amirkhani, Mohammad Taher Pilehvar
2023 J jnl
RET-LLM: Towards a General Read-Write Memory for Large Language Models.
CoRR
Ali Modarressi, Ayyoob Imani, Mohsen Fayyaz, Hinrich Schütze
2022 — conf
AdapLeR: Speeding up Inference by Adaptive Length Reduction.
ACL (1)
Ali Modarressi, Hosein Mohebbi, Mohammad Taher Pilehvar
2022 J jnl
AdapLeR: Speeding up Inference by Adaptive Length Reduction.
CoRR
Ali Modarressi, Hosein Mohebbi, Mohammad Taher Pilehvar
2022 J jnl
BERT on a Data Diet: Finding Important Examples by Gradient-Based Pruning.
CoRR
Mohsen Fayyaz, Ehsan Aghazadeh, Ali Modarressi, Mohammad Taher Pilehvar, Yadollah Yaghoobzadeh, Samira Ebrahimi Kahou
2022 — conf
GlobEnc: Quantifying Global Token Attribution by Incorporating the Whole Encoder Layer in Transformers.
NAACL-HLT
Ali Modarressi, Mohsen Fayyaz, Yadollah Yaghoobzadeh, Mohammad Taher Pilehvar
2022 J jnl
GlobEnc: Quantifying Global Token Attribution by Incorporating the Whole Encoder Layer in Transformers.
CoRR
Ali Modarressi, Mohsen Fayyaz, Yadollah Yaghoobzadeh, Mohammad Taher Pilehvar
2021 — conf
Exploring the Role of BERT Token Representations to Explain Sentence Probing Results.
EMNLP (1)
Hosein Mohebbi, Ali Modarressi, Mohammad Taher Pilehvar
2021 J jnl
Exploring the Role of BERT Token Representations to Explain Sentence Probing Results.
CoRR
Hosein Mohebbi, Ali Modarressi, Mohammad Taher Pilehvar
2021 — conf
Not All Models Localize Linguistic Knowledge in the Same Place: A Layer-wise Probing on BERToids' Representations.
BlackboxNLP@EMNLP
Mohsen Fayyaz, Ehsan Aghazadeh, Ali Modarressi, Hosein Mohebbi, Mohammad Taher Pilehvar
2021 J jnl
Not All Models Localize Linguistic Knowledge in the Same Place: A Layer-wise Probing on BERToids' Representations.
CoRR
Mohsen Fayyaz, Ehsan Aghazadeh, Ali Modarressi, Hosein Mohebbi, Mohammad Taher Pilehvar