Publications

2026

1. ACL

   

   Truth as a Trajectory: What Internal Representations Reveal About Large Language Model Reasoning

   Hamed Damirchi, Ignacio Meza De la Jara, Ehsan Abbasnejad, and 3 more authors

   In Annual Meeting of the Association for Computational Linguistics (ACL) , 2026

   Spotlight, ACL 2026 main conference

   arXiv Cite PDF

   @inproceedings{De_la_Jara_2026_ACL,
     author = {Damirchi, Hamed and De la Jara, Ignacio Meza and Abbasnejad, Ehsan and Shamsi, Afshar and Zhang, Zhen and Shi, Javen},
     title = {Truth as a Trajectory: What Internal Representations Reveal About Large Language Model Reasoning},
     booktitle = {Annual Meeting of the Association for Computational Linguistics (ACL)},
     year = {2026},
     note = {Spotlight, ACL 2026 main conference},
   }

2. arXiv

   

   Level, Sharpness, and Corpus: Why Zero-Shot OOD Detector Rankings Do Not Transfer

   Ignacio M. De la Jara, Cristian Rodriguez-Opazo, Stephen Gould, and 1 more author

   arXiv preprint , 2026

   A portability audit across seventeen datasets, three vision-language models and seven detectors, showing that zero-shot OOD detector rankings reverse across deployments.

   Abs arXiv Cite

   Selecting a zero-shot out-of-distribution (OOD) detector for a new deployment is typically based on benchmark rankings, implicitly assuming that the highest-ranked detector will transfer across domains. We show that this assumption does not hold. Through a controlled portability audit across seventeen in-distribution datasets, three vision-language models, and seven representative zero-shot OOD detectors, we find that detector rankings reverse across deployments, every detector exceeds 80% FPR95 on at least one domain, and the preferred detector depends on both the in-distribution data and the underlying VLM. We trace these reversals to complementary evidence channels in vision-language logits. Corpus-free detectors rely on different combinations of absolute match level and relative or spatial sharpness, while WordNet-based methods additionally depend on external semantic coverage. A simple proposition shows that level and sharpness cannot generally be recovered from one another, explaining why no single detector transfers reliably across deployments. Motivated by this diagnosis, we introduce the Complementary Evidence Guard (CEG), a detector-agnostic wrapper that preserves complementary evidence through a non-compensatory fusion of the base detector, level, and sharpness using only empirical in-distribution percentiles. Controls replacing these channels with entropy, logit variance, or random noise do not reproduce the gains. Without OOD samples, auxiliary corpora, or learned fusion, CEG reduces detector sensitivity and improves GL-MCM from 38.1 to 28.8 and MCM from 42.6 to 30.5 family-balanced FPR95.

   @misc{De_la_Jara_2026_arXiv_Level_Sharpness_Corpus,
     author = {De la Jara, Ignacio M. and Rodriguez-Opazo, Cristian and Gould, Stephen and Ranasinghe, Damith},
     title = {Level, Sharpness, and Corpus: Why Zero-Shot OOD Detector Rankings Do Not Transfer},
     year = {2026},
     eprint = {2607.26582},
     archiveprefix = {arXiv},
     primaryclass = {cs.CV},
     url = {https://arxiv.org/abs/2607.26582},
   }

3. arXiv

   

   Representation Trajectories Matters: Complementary Evidence for OOD Detection and Image Classification

   Ignacio M. De la Jara, Cristian Rodriguez-Opazo, Hamed Damirchi, and 2 more authors

   arXiv preprint , 2026

   Keeps sample identity across depth to study the transformations connecting successive representations, and shows the resulting computation path carries evidence the final representation discards.

   Abs arXiv Cite

   Vision models do not form a representation at once; each block revises it. We ask whether the resulting computation path contains evidence that the final representation discards, and whether that evidence improves OOD detection and image classification on clean and shifted data. Unlike approaches that treat intermediate layers as separate snapshots, we retain sample identity across depth and study the transformations connecting successive states. We separate class-coherent transport from input-specific innovation, and coordinate movement from relational reorganization. Across supervised, self-supervised, vision–language, hierarchical, and convolutional encoders, these paths show strong sample-specific continuity and architecture-specific depth profiles that recur across datasets. They are also practically useful. An ID-only transition-surprise score complements strong final-state detectors, reducing FPR95 in 131/152 non-saturated comparisons on a balanced OpenOOD grid; gains are largest for visually disruptive and semantically far shifts, and remain positive on near-OOD for most detectors. Frozen update probes improve 71/72 clean model–dataset cases, while shifted-data gains vary with architecture and corruption type. Computation paths therefore provide a broadly useful reliability signal whose value is determined jointly by model organization and the shift encountered.

   @misc{De_la_Jara_2026_arXiv_Representation_Trajectories,
     author = {De la Jara, Ignacio M. and Rodriguez-Opazo, Cristian and Damirchi, Hamed and Gould, Stephen and Ranasinghe, Damith},
     title = {Representation Trajectories Matters: Complementary Evidence for OOD Detection and Image Classification},
     year = {2026},
     eprint = {2607.26565},
     archiveprefix = {arXiv},
     primaryclass = {cs.CV},
     url = {https://arxiv.org/abs/2607.26565},
   }

2025

1. NeurIPS

   

   Mysteries of the Deep: Role of Intermediate Representations in Out-of-Distribution Detection

   Ignacio Meza De la Jara, Cristian Rodriguez-Opazo, Damien Teney, and 2 more authors

   In Advances in Neural Information Processing Systems , 2025

   arXiv Cite PDF Code Website

   @inproceedings{De_la_Jara_2025_NeurIPS,
     author = {De la Jara, Ignacio Meza and Rodriguez-Opazo, Cristian and Teney, Damien and Ranasinghe, Damith and Abbasnejad, Ehsan},
     title = {Mysteries of the Deep: Role of Intermediate Representations in Out-of-Distribution Detection},
     booktitle = {Advances in Neural Information Processing Systems},
     year = {2025},
   }

2023

1. ICCVW

   

   An Empirical Study of the Effect of Video Encoders on Temporal Video Grounding

   Ignacio M. De la Jara, Cristian Rodriguez-Opazo, Edison Marrese-Taylor, and 1 more author

   In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) Workshops , Oct 2023

   arXiv Cite PDF

   @inproceedings{De_la_Jara_2023_ICCV,
     author = {De la Jara, Ignacio M. and Rodriguez-Opazo, Cristian and Marrese-Taylor, Edison and Bravo-Marquez, Felipe},
     title = {An Empirical Study of the Effect of Video Encoders on Temporal Video Grounding},
     booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) Workshops},
     month = oct,
     year = {2023},
     pages = {2850-2855},
   }