Interesting papers
All published papers whose full text was read, ranked by interest score.
| # | Score | Paper | Year |
|---|---|---|---|
| 26 | 7.5 |
Check The Scoreboard: An Analysis of Scoring Schemes on Multiple-Choice Evaluation
arXiv · doi:10.48550/arxiv.2608.29887 Novelty 8.0 · Rigor 8.0 · Engineering 8.0 · Consensus 5.0 · Topic 8.0 |
2026 |
| 27 | 7.5 |
NeutronGym: Physics-Graded Neutron Instrument Design for LLM Agents
arXiv · doi:10.48550/arxiv.2610.03631 Novelty 8.0 · Rigor 8.0 · Engineering 8.0 · Consensus 5.0 · Topic 8.0 |
2026 |
| 28 | 7.5 |
Proxy Confidence: Auditing Black-Box LLM Agents with a Surrogate's Log-Probabilities
arXiv · doi:10.48550/arxiv.2610.03894 Novelty 8.0 · Rigor 8.0 · Engineering 8.0 · Consensus 5.0 · Topic 8.0 |
2026 |
| 29 | 7.5 |
Pressure, Context, and Machine Self-Control: A Criminological Test of Reward Hacking in Generative AI Models
arXiv · doi:10.48550/arxiv.2610.04793 Novelty 8.0 · Rigor 8.0 · Engineering 8.0 · Consensus 5.0 · Topic 9.0 |
2026 |
| 30 | 7.5 |
Synthetic Worldview Reconstruction (SWR): A Methodological Framework for LLM-Assisted Group-Level Belief System Analysis
Zenodo (CERN European Organization for Nuclear Research) · doi:10.5281/zenodo.18736720 Novelty 8.0 · Rigor 8.0 · Engineering 8.0 · Consensus 5.0 · Topic 7.0 |
2026 |
| 31 | 7.5 |
From Feature-Based Models to Generative AI: Validity Evidence for Constructed Response Scoring
arXiv · doi:10.48550/arxiv.2603.19280 Novelty 7.0 · Rigor 8.0 · Engineering 7.0 · Consensus 7.5 · Topic 8.5 |
2026 |
| 32 | 7.5 |
An Interpretable and Scalable Framework for Evaluating Large Language Models
arXiv · doi:10.48550/arxiv.2605.07046 Novelty 7.5 · Rigor 8.5 · Engineering 8.0 · Consensus 6.5 · Topic 7.5 |
2026 |
| 33 | 7.5 |
Are Near-Tied LLM Rankings Robust to Family-DIF-Guided Benchmark Recomposition?
arXiv · doi:10.48550/arxiv.2609.00482 Novelty 8.0 · Rigor 8.0 · Engineering 7.0 · Consensus 6.0 · Topic 7.0 |
2026 |
| 34 | 7.5 |
Evaluating LLM-as-a-Judge Beyond Score Alignment: A Psychometric Analysis of Residual Judging Difficulty
arXiv · doi:10.48550/arxiv.2610.02877 Novelty 8.0 · Rigor 8.0 · Engineering 7.0 · Consensus 6.0 · Topic 9.0 |
2026 |
| 35 | 7.4 |
Interpretable, Fairly Evaluated Automated L2 Speaking Assessment that Beats the Single-Human Ceiling and Why Pause Encoding Does Not Change LLM Fluency Scores
arXiv · doi:10.48550/arxiv.2608.26137 Novelty 6.5 · Rigor 8.0 · Engineering 6.5 · Consensus 7.5 · Topic 7.5 |
2026 |
| 36 | 7.4 |
Machine--human agreement in marking undergraduate assignments is specific to model, prompt and institution
arXiv · doi:10.48550/arxiv.2610.18019 Novelty 7.5 · Rigor 8.5 · Engineering 5.0 · Consensus 8.0 · Topic 8.5 |
2026 |
| 37 | 7.4 |
Inside the Scorer: Mechanistic Evidence on Rubric Control in LLM-Based Automated Essay Scoring from 25 Base Models
arXiv · doi:10.48550/arxiv.2610.25724 Novelty 8.0 · Rigor 8.5 · Engineering 5.0 · Consensus 8.0 · Topic 8.5 |
2026 |
| 38 | 7.4 |
MisconceptDialogue: Evaluation Criteria Shape Measured Success in LLM Math Tutoring
arXiv (Cornell University) · doi:10.48550/arxiv.2610.31679 Novelty 7.5 · Rigor 8.0 · Engineering 6.5 · Consensus 7.5 · Topic 7.5 |
2026 |
| 39 | 7.4 |
What exam scores can and cannot prove about unauthorized AI assistance: Evidence from a highly public classroom episode
arXiv · doi:10.48550/arxiv.2607.27978 Novelty 8.0 · Rigor 8.0 · Engineering 5.0 · Consensus 8.0 · Topic 9.0 |
2026 |
| 40 | 7.4 |
Scalable Text-Embedding-informed Cognitive Diagnosis of Large Language Models
arXiv · doi:10.48550/arxiv.2603.14676 Novelty 8.0 · Rigor 8.5 · Engineering 7.5 · Consensus 6.5 · Topic 7.0 |
2026 |
| 41 | 7.3 |
Spatial-Sign based High dimensional Change Point Inference
arXiv · doi:10.48550/arxiv.2504.19306 Novelty 8.0 · Rigor 8.0 · Engineering 7.0 · Consensus 5.0 · Topic 7.0 |
2025 |
| 42 | 7.3 |
Identifying Causes of Test Unfairness: Manipulability and Separability
arXiv · doi:10.48550/arxiv.2601.13449 Novelty 8.0 · Rigor 8.0 · Engineering 7.0 · Consensus 5.0 · Topic 8.0 |
2026 |
| 43 | 7.3 |
What Makes Something Hard(er)? Explaining Question Difficulty in Natural Language
arXiv · doi:10.48550/arxiv.2610.01627 Novelty 8.0 · Rigor 8.0 · Engineering 7.0 · Consensus 5.0 · Topic 8.0 |
2026 |
| 44 | 7.3 |
Representational Control over Self-Report & Behavior Coherence in LLM Risk-Taking
arXiv · doi:10.48550/arxiv.2610.04125 Novelty 8.0 · Rigor 8.0 · Engineering 7.0 · Consensus 5.0 · Topic 8.0 |
2026 |
| 45 | 7.3 |
Auditing LLM Benchmarks with Item Response Theory
arXiv · doi:10.48550/arxiv.2605.30504 Novelty 8.0 · Rigor 8.0 · Engineering 7.0 · Consensus 7.0 · Topic 8.0 |
2026 |
| 46 | 7.3 |
From Scoring to Explanations: Evaluating SHAP and LLM Rationales for Rubric-based Teaching Quality Assessment
arXiv · doi:10.48550/arxiv.2606.05180 Novelty 7.0 · Rigor 8.0 · Engineering 7.0 · Consensus 7.0 · Topic 7.0 |
2026 |
| 47 | 7.3 |
Prompt Sensitivity in LLM-Based Essay Scoring is Model-Specific
CLOK (University of Central Lancashire) · doi:10.5281/zenodo.21039771 Novelty 7.5 · Rigor 7.0 · Engineering 6.5 · Consensus 8.0 · Topic 8.5 |
2026 |
| 48 | 7.3 |
Algorithmic drift in generative adaptive learning
arXiv · doi:10.48550/arxiv.2610.31041 Novelty 7.5 · Rigor 7.5 · Engineering 7.5 · Consensus 7.0 · Topic 7.5 |
2026 |
| 49 | 7.3 |
Which AI model is truly the best: how reliable are public benchmarks of frontier AI models?
Zenodo (CERN European Organization for Nuclear Research) · doi:10.5281/zenodo.22998204 Novelty 7.0 · Rigor 8.0 · Engineering 6.5 · Consensus 8.5 · Topic 6.5 |
2026 |
| 50 | 7.2 |
A regularised latent‐class item response model for detecting measurement non‐invariance in ordinal response scales
British Journal of Mathematical and Statistical Psychology · doi:10.1111/bmsp.70076 Novelty 7.0 · Rigor 8.0 · Engineering 8.0 · Consensus 5.0 · Topic 8.0 |
2026 |