Researchers have developed ATTRICITE, a new 4-billion parameter open-source model designed to improve faithful citation attribution in scientific literature. The model focuses on citation recovery, identifying the specific paper cited by an author from a given passage. ATTRICITE was trained using the CITEALIGN dataset and fine-tuned with GRPO, achieving a target-match accuracy of 59.8%, which surpasses larger models like gpt-oss-20b and approaches GPT-5.4-mini. AI
IMPACT This research could lead to more reliable AI systems for scientific literature analysis and knowledge discovery.
RANK_REASON The item describes a new open-source model and dataset released for research purposes, detailing its training and performance on a specific task. [lever_c_demoted from research: ic=1 ai=1.0]
- alphaXiv
- ATTRICITE
- CatalyzeX Code Finder for Papers
- CITEALIGN
- CiteGuard
- Connected Papers
- CORE Recommender
- DagsHub
- Gemma 4 31B IT
- Gotit.pub
- GPT 5.4 Mini
- gpt-oss-20b
- GRPO
- Hugging Face
- Influence Flower
- Litmaps
- Qwen3-4B
- ScienceCast
- scite Smart Citations
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →