ServiceNow AI Research

Large Language Models

Grounding Computer Use Agents on Human Demonstrations
Grounding Computer Use Agents on Human Demonstrations
Building reliable computer-use agents requires grounding: accurately connecting natural language instructions to the correct on-screen …
Aarash Feizi, Shravan Nayak, Xiangru Jian, Kevin Qinghong Lin, Kaixin Li, Rabiul Awal, Xing Han Lu, Johan Obando, Juan A. Rodriguez, Nicolas Chapados, David Vazquez, Adriana Romero Soriano, Reihaneh Rabbany, Perouz Taslakian, Christopher Pal, Spandana Gella, Sai Rajeswar Mudumba
International Conference on Learning Representations,  2026.
Paper Cite Code
Causal Differentiating Concepts: Interpreting LM Behavior via Causal Representation Learning
Causal Differentiating Concepts: Interpreting LM Behavior via Causal Representation Learning
Language model activations entangle concepts that mediate their behavior, making it difficult to interpret these factors, which has …
Navita Goyal, Hal Daumé III, Alexandre Drouin, Dhanya Sridhar
Neural Information Processing Systems (NeurIPS),  2025.
Paper Cite
Using Scaling Laws for Data Source Utility Estimation in Domain-Specific Pre-Training
Using Scaling Laws for Data Source Utility Estimation in Domain-Specific Pre-Training
We introduce a framework for optimizing domain-specific dataset construction in foundation model training. Specifically, we seek a …
Oleksiy Ostapenko, Charles Guille-Escuret, Luke Kumar, Max Tian, Denis Kocetkov, Gopeshh Subbaraj, Raymond Li, Joel Lamy Poirier, Sébastien Paquet, Torsten Scholak
Conference on Language Modeling Workshops,  2025.
Paper Cite
BiXSE: Improving Dense Retrieval via Probabilistic Graded Relevance Distillation
BiXSE: Improving Dense Retrieval via Probabilistic Graded Relevance Distillation

Neural sentence embedding models for dense retrieval typically rely on binary relevance labels, treating query-document pairs as …

Christos Tsirigotis, Vaibhav Adlakha, João Monteiro, Aaron Courville, Perouz Taslakian
Conference on Language Modeling (COLM),  2025.
Paper Cite Code
Unifying Autoregressive and Diffusion-Based Sequence Generation
Unifying Autoregressive and Diffusion-Based Sequence Generation
We present significant extensions to diffusion-based sequence generation models, blurring the line with autoregressive language models. …
Nima Fathi, Torsten Scholak, Pierre-André Noël
Conference on Language Modeling (COLM),  2025.
Paper Cite Code Video
FocusAgent: Simple Yet Effective Ways of Trimming the Large Context of Web Agents
FocusAgent: Simple Yet Effective Ways of Trimming the Large Context of Web Agents
Web agents powered by large language models (LLMs) must process lengthy web page observations to complete user goals; these pages often …
Imene Kerboua, Sahar Omidi Shayegan, Xing Han Lu, Léo Boisvert, Megh Thakkar, Massimo Caccia, Jérémy Espinas, Alex Aussem, Véronique Eglin, Alexandre Lacoste
Transactions on Machine Learning Research (TMLR),  2025.
Paper Cite Code
LLMs can learn self-restraint through iterative self-reflection
LLMs can learn self-restraint through iterative self-reflection
In order to be deployed safely, Large Language Models (LLMs) must be capable of dynamically adapting their behavior based on their …
Alexandre Piche, Aristides Milios, Dzmitry Bahdanau, Christopher Pal
Transactions on Machine Learning Research (TMLR),  2025.
Paper Cite Video
Auto-Cypher: Improving LLMs on Cypher generation via LLM-supervised generation-verification framework
Auto-Cypher: Improving LLMs on Cypher generation via LLM-supervised generation-verification framework
Graph databases like Neo4j are gaining popularity for handling complex, interconnected data, over traditional relational databases in …
Aman Tiwari, Shiva Krishna Reddy Malay, Vikas Yadav, Masoud Hashemi, Sathwik Tejaswi Madhusudhan
North American Chapter of the Association for Computational Linguistics (NAACL),  2025.
Paper Cite
M2Lingual: Enhancing Multilingual, Multi-Turn Instruction Alignment in Large Language Models
M2Lingual: Enhancing Multilingual, Multi-Turn Instruction Alignment in Large Language Models
Instruction finetuning (IFT) is critical for aligning Large Language Models (LLMs) to follow instructions. While many effective IFT …
Rishabh Maheshwary, Vikas Yadav, Hoang Nguyen, Khyati Mahajan, Sathwik Tejaswi Madhusudhan
North American Chapter of the Association for Computational Linguistics (NAACL),  2025.
Paper Cite Code
Prompting with Phonemes: Enhancing LLM Multilinguality for non-Latin Script Languages
Prompting with Phonemes: Enhancing LLM Multilinguality for non-Latin Script Languages
Multilingual LLMs have achieved remarkable benchmark performance, but we find they continue to underperform on non-Latin script …
Hoang Nguyen, Khyati Mahajan, Vikas Yadav, Philip S. Yu, Masoud Hashemi, Rishabh Maheshwary
North American Chapter of the Association for Computational Linguistics (NAACL),  2025.
Paper Cite