Introduces EURO-5K dataset from 136 EU acts and benchmarks full fine-tuning vs QLoRA for BERT and LLM models on reporting obligation extraction, reporting 0.89 F1 with limited gains from legal pretraining except under parameter-efficient adaptation.
hub
S ci BERT : A pretrained language model for scientific text
17 Pith papers cite this work, alongside 3,014 external citations. Polarity classification is still indexing.
17
Pith papers citing it
3,014
external citations · OpenAlex
hub tools
citation-role summary
background 2
citation-polarity summary
roles
background 2polarities
background 2representative citing papers
RISE is an inference-time semantic reranking framework that refines low-confidence predictions in rhetorical role labeling using contrastively learned label representations, delivering an average +9.15 macro-F1 gain on hard examples across eight datasets and seven models.
SciNLP is the first full-text entity and relation extraction benchmark for the NLP domain, built from 60 manually annotated publications and used to evaluate models and construct a domain knowledge graph.
Evaluation of 22 LLMs shows they are more susceptible to spin in medical abstracts than humans but can recognize and mitigate it when prompted.
RWGBench measures related-work generation by citation choices, and shows citation-focused metrics expose failures that text-similarity and LLM-judge scores miss.
Vocabulary adaptation via targeted token addition and replacement improves semantic similarity, domain word usage, and training efficiency for LLM summarization in legal and medical domains.
SPARK constructs unified knowledge graphs from multi-document scientific literature to ground self-play RL with asymmetric roles and verifiable rewards, outperforming flat-corpus baselines especially on longer-hop reasoning tasks.
Matrix factorization on a literature-mined concept-object graph predicts future associations in astronomy better than neighborhood similarity or recency heuristics.
IDRBench, a new benchmark, shows that LLMs can produce plausible interdisciplinary research ideas but cannot reliably identify or distinguish genuine cross-disciplinary integration.
Continued pretraining of Code Llama on Proof-Pile-2 yields Llemma, an open math-specialized LLM that beats known open base models on MATH and supports tool use plus formal proving out of the box.
BloombergGPT is a 50B parameter LLM trained on a 708B token mixed financial and general dataset that outperforms prior models on financial benchmarks while preserving general LLM performance.
Applies pretrained deep learning models with data augmentation to classify algorithm mention motivations in NLP papers, reporting that direct use dominates and motivation diversity has increased over time.
Entity extraction and z-score analysis of NLP papers shows pre-trained models like BERT and Transformer as mainstream with accelerating acceptance of new high-impact technologies.
Compass is an expert-guided LLM agent framework that extracts 3,751 marine Pb records from 230k papers to build the largest integrated database, achieving 92% accuracy via multi-layered validation.
The hybrid method with LLM-augmented data achieves F1 improvements of 7-24 points over baselines on five Vietnamese domain datasets.
Numerical scores predict ICLR acceptance at 91% accuracy while review text reaches only 81%, because politeness makes rejected papers' reviews contain more positive than negative words.
Further pre-training ModernBERT on US court opinions improves results on legal datasets compared to the base model, with gains similar to early BERT domain adaptation work.
citing papers explorer
-
EURO-5K: When Does Domain Pretraining Matter? Benchmarking Transformers for EU Reporting Obligation Extraction
Introduces EURO-5K dataset from 136 EU acts and benchmarks full fine-tuning vs QLoRA for BERT and LLM models on reporting obligation extraction, reporting 0.89 F1 with limited gains from legal pretraining except under parameter-efficient adaptation.
-
Semantic Reranking at Inference Time for Hard Examples in Rhetorical Role Labeling
RISE is an inference-time semantic reranking framework that refines low-confidence predictions in rhetorical role labeling using contrastively learned label representations, delivering an average +9.15 macro-F1 gain on hard examples across eight datasets and seven models.
-
SciNLP: A Domain-Specific Benchmark for Full-Text Scientific Entity and Relation Extraction in NLP
SciNLP is the first full-text entity and relation extraction benchmark for the NLP domain, built from 60 manually annotated publications and used to evaluate models and construct a domain knowledge graph.
-
Caught in the Web of Words: Do LLMs Fall for Spin in Medical Literature?
Evaluation of 22 LLMs shows they are more susceptible to spin in medical abstracts than humans but can recognize and mitigate it when prompted.
-
RWGBench: Evaluating Scholarly Positioning in Related Work Generation
RWGBench measures related-work generation by citation choices, and shows citation-focused metrics expose failures that text-similarity and LLM-judge scores miss.
-
Learning Faster with Better Tokens: Parameter-Efficient Vocabulary Adaptation for Specialized Text Summarization
Vocabulary adaptation via targeted token addition and replacement improves semantic similarity, domain word usage, and training efficiency for LLM summarization in legal and medical domains.
-
SPARK: Self-Play with Asymmetric Reward from Knowledge Graphs
SPARK constructs unified knowledge graphs from multi-document scientific literature to ground self-play RL with asymmetric roles and verifiable rewards, outperforming flat-corpus baselines especially on longer-hop reasoning tasks.
-
Predicting New Concept-Object Associations in Astronomy by Mining the Literature
Matrix factorization on a literature-mined concept-object graph predicts future associations in astronomy better than neighborhood similarity or recency heuristics.
-
IDRBench: Understanding the Capability of Large Language Models on Interdisciplinary Research
IDRBench, a new benchmark, shows that LLMs can produce plausible interdisciplinary research ideas but cannot reliably identify or distinguish genuine cross-disciplinary integration.
-
Llemma: An Open Language Model For Mathematics
Continued pretraining of Code Llama on Proof-Pile-2 yields Llemma, an open math-specialized LLM that beats known open base models on MATH and supports tool use plus formal proving out of the box.
-
BloombergGPT: A Large Language Model for Finance
BloombergGPT is a 50B parameter LLM trained on a 708B token mixed financial and general dataset that outperforms prior models on financial benchmarks while preserving general LLM performance.
-
Exploring Motivations for Algorithm Mention in the Domain of Natural Language Processing: A Deep Learning Approach
Applies pretrained deep learning models with data augmentation to classify algorithm mention motivations in NLP papers, reporting that direct use dominates and motivation diversity has increased over time.
-
Revealing the Technology Development of Natural Language Processing: A Scientific Entity-Centric Perspective
Entity extraction and z-score analysis of NLP papers shows pre-trained models like BERT and Transformer as mainstream with accelerating acceptance of new high-impact technologies.
-
Compass: Navigating Global Marine Lead Data Integration through Expert-Guided LLM Agent
Compass is an expert-guided LLM agent framework that extracts 3,751 marine Pb records from 230k papers to build the largest integrated database, achieving 92% accuracy via multi-layered validation.
-
A Hybrid Method for Low-Resource Named Entity Recognition
The hybrid method with LLM-augmented data achieves F1 improvements of 7-24 points over baselines on five Vietnamese domain datasets.
-
Decoupling Scores and Text: The Politeness Principle in Peer Review
Numerical scores predict ICLR acceptance at 91% accuracy while review text reaches only 81%, because politeness makes rejected papers' reviews contain more positive than negative words.
-
Legal Domain Adaptation of Modern BERT Models
Further pre-training ModernBERT on US court opinions improves results on legal datasets compared to the base model, with gains similar to early BERT domain adaptation work.