Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
May 14, 2026Intelligent Data Analysis

IN-LAND: Improving large language models’ performance on abbreviation disambiguation

View Full Paper
Ask AI
Bookmark
Share

Authors

NLNankai LinGuangdong University of Foreign StudiesBRBowen RuanGuangdong University of Foreign StudiesHBHaoyuan BuGuangdong University of Foreign Studies

Discussion

Loading...

Member takes

Implication

Randomized trial evaluates the IN-LAND framework for abbreviation disambiguation in various languages, indicating notable improvements.

Key Points

  • This research aims to enhance abbreviation disambiguation performance in large language models by developing the IN-LAND framework.
  • Developed the IN-LAND framework for abbreviation disambiguation using a novel selection mechanism for demonstration examples.
  • Evaluated performance across four large language models in languages like French, Legal English, and Malay.
  • Compared IN-LAND results to a Random K-shot in-context learning baseline.
  • In Malay, average macro F1 score improved by 7.11% and accuracy by 9.19%.
  • In French, the macro F1 score increased by 2.69% and accuracy by 3.17%.
  • Legal English showed a 1.04% improvement in average accuracy while maintaining competitive macro F1 scores.

Cite This Study

Lin et al. (2026) studied this question.

synapsesocial.com/papers/6a05685ca550a87e60a20ec8https://doi.org/10.1177/1088467x261444141
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Dr.ICL: Demonstration-Retrieved In-context Learning2024 · 11 citations
  2. 2What Makes Good In-Context Examples for GPT-3?2022 · 547 citations
  3. 3Fantastically Ordered Prompts and Where to Find Them: Overcoming Few-Shot Prompt Order Sensitivity2022 · 491 citations
  4. 4ProGen: Progressive Zero-shot Dataset Generation via In-context Feedback2022 · 24 citations
  5. 5TriviaQA: A Large Scale Distantly Supervised Challenge Dataset for Reading Comprehension2017 · 1,697 citations