Research map / Large language models / Language, safety and interpretability
Decoding Open-Ended Information Seeking Goals from Eye Movements in Reading
Cfir Avraham Hadar, Omer Shubi, Yoav Meiri, Amit Heshes, Yevgeni Berzak
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use RoBERTEye-Fixations instead of Haller RNN | goal selection model · OneStop Same Span | selection accuracy (%): 52.1 -> 57.3 | Improved |
| use RoBERTEye-Fixations instead of Haller RNN | goal selection model · OneStop All | selection accuracy (%): 41.8 -> 49.3 | Improved |
| use RoBERTEye-Fixations instead of Haller RNN | goal selection model · OneStop Different Spans | selection accuracy (%): 65.6 -> 70.9 | Improved |
| use reading-time weighted passage embedding similarity instead of chance | heuristic baseline · OneStop All | selection accuracy (%): 33.0 -> 33.4 | No clear effect |
| use RT similarity to question-word similarities instead of chance | heuristic baseline · OneStop All | selection accuracy (%): 33.0 -> 34.2 | No clear effect |
| longer per-word reading time within the critical span | RoBERTEye-Fixations · OneStop trial features | probability of correct question | Improved |
| longer per-word reading time before and after the critical span | RoBERTEye-Fixations · OneStop trial features | probability of correct question | Hurt |
| longer paragraphs | RoBERTEye-Fixations · OneStop trial features | probability of correct question | Improved |
| lexical overlap of question with critical span | RoBERTEye-Fixations · OneStop trial features | probability of correct question | No clear effect |
| paragraph difficulty (original vs simplified) | RoBERTEye-Fixations · OneStop trial features | probability of correct question | No clear effect |
| add eye movements to Llama 3.1 fine-tuned question reconstruction (DalEye-Llama) | Llama 3.1 · OneStop New Text reconstruction | question word / category | Mixed |
Closest papers by meaning
- Head Pursuit: Probing Attention Specialization in Multimodal Transformers (NeurIPS 2025 · similarity 0.83)
- Reading Recognition in the Wild (NeurIPS 2025 · similarity 0.83)
- Coarse-to-Fine Highlighting: Reducing Knowledge Hallucination in Large Language Models (ICML 2024 · similarity 0.83)
- MCIF: Multimodal Crosslingual Instruction-Following Benchmark from Scientific Talks (ICLR 2026 · similarity 0.82)
- EyEar: Learning Audio Synchronized Human Gaze Trajectory Based on Physics-Informed Dynamics (AAAI 2025 · similarity 0.82)
- MMIE: Massive Multimodal Interleaved Comprehension Benchmark for Large Vision-Language Models (ICLR 2025 · similarity 0.82)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
