arXiv ·
SWE-Duel: Towards a Self-Scaling Arena for Coding Agents
Public paper from arXiv. Open the source to read the full record.
Home / Browse
Models & research
Examine models, papers and evaluations with their methods and limitations.
802 public stories
arXiv ·
Contextual Correction of Parallel Draftsfor Speculative Decoding
Public paper from arXiv. Open the source to read the full record.
arXiv ·
CLUE: A Contrastive Learning Framework for Unsupervised Event Type Induction and Annotation
Public paper from arXiv. Open the source to read the full record.
arXiv ·
One Model to Loop Them All: Elastic Latent Recurrence for Language Models
Public paper from arXiv. Open the source to read the full record.
arXiv ·
Reflective Prompt Optimization of a Deep-Research Agent: A Topic-Disjoint, Multi-Seed Protocol for Measuring Transferable Gains and Prompt-Level Overfitting
Public paper from arXiv. Open the source to read the full record.
arXiv ·
Self-Preservation in Language Models
Public paper from arXiv. Open the source to read the full record.
arXiv ·
GeopoliticalQA: Benchmarking Retrieval and Generation over Intelligence Reports
Public paper from arXiv. Open the source to read the full record.
arXiv ·
perturber: A Stateless Service for Semantics-Preserving Text Perturbations
Public paper from arXiv. Open the source to read the full record.
arXiv ·
GeoFitBench-LLM: A Benchmark for Assessing Large Language Models' Capabilities in Geospatial Data Requirements Elicitation for Geo-Analytic Tasks
Public paper from arXiv. Open the source to read the full record.
arXiv ·
Activating the Annotator: Operationalising Lived Experience in Data Annotation
Public paper from arXiv. Open the source to read the full record.
arXiv ·
Unsupervised Multimodal Reward Modeling
Public paper from arXiv. Open the source to read the full record.
arXiv ·
U4RASD at KnowledgeGraphEval-2026: Ontology-Constrained Encoder and LLM Systems for Arabic Relation Extraction
Public paper from arXiv. Open the source to read the full record.
arXiv ·
MADS: Multi-Agent Dialog Simulation for Robust Conversational Agents
Public paper from arXiv. Open the source to read the full record.
arXiv ·
DistressBench: Evaluating Multi-Dimensional Crisis Support in Large Language Models
Public paper from arXiv. Open the source to read the full record.
arXiv ·
People Making Conversation: Evaluating How Naturally Speech Models Converse
Public paper from arXiv. Open the source to read the full record.
arXiv ·
When Charts Deceive Web Agents: A Paired Benchmark for Misleading Visualization-Grounded Tasks
Public paper from arXiv. Open the source to read the full record.
arXiv ·
COFFEQA: Counterfactual-Guided Few-Shot Distillation for LLM Question Answering
Public paper from arXiv. Open the source to read the full record.
arXiv ·
Fear Is Not Always Fear: Variation in Emotion Label Equivalence Across Datasets
Public paper from arXiv. Open the source to read the full record.
arXiv ·
Unreliable Narrator Framing as a Cost-effective Mechanism for Satire Explanation of Australian and American Satirical News
Public paper from arXiv. Open the source to read the full record.
arXiv ·
Time Delay System for Sequence Modeling
Public paper from arXiv. Open the source to read the full record.
arXiv ·
Temporal Evidence Forests for Long-Horizon Multi-Party Conversational Memory
Public paper from arXiv. Open the source to read the full record.
arXiv ·
Rethinking the Roles of Diffusion and Decoding: Predictive-State Flows for Parallel Language Generation
Public paper from arXiv. Open the source to read the full record.
arXiv ·
SEEDS: Preview-Guided Evidence Admission for Search Agents
Public paper from arXiv. Open the source to read the full record.
arXiv ·
Surfing the Activation Wave: Analyzing the Information and Geometry of LLM Hidden States for Speculative Decoding
Public paper from arXiv. Open the source to read the full record.
arXiv ·
Forecasting Corpus-Sampling Disagreement in Masked Text Reconstruction
Public paper from arXiv. Open the source to read the full record.
arXiv ·
Language Models Have Interpretable Narrative Mechanisms
Public paper from arXiv. Open the source to read the full record.
arXiv ·
What Should SFT Learn? Conditional Information Gain for Generalizable Post-Training
Public paper from arXiv. Open the source to read the full record.
arXiv ·
Evaluating LLM Creativity through Long-Tail Performance
Public paper from arXiv. Open the source to read the full record.
arXiv ·
Evo$^2$: Evolving the Evolver through Weight-Harness Co-Evolution
Public paper from arXiv. Open the source to read the full record.
V2EX ·
[程序员] 和同事吹水吹出来的：用三国志的数值给大模型排了个座位
Public article from V2EX. Open the source to read the full record.