arXiv ·
From Copying to Composition: How Language Models Develop In-Context Learning
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 ·
DecisionRobust: Benchmarking Robustness for JEV Style Decision Models Across Text and Vision
Public paper from arXiv. Open the source to read the full record.
arXiv ·
CAFE: Confining Feature Distillation to Language-Routed Adapters for Compact Multilingual Automatic Speech Recognition
Public paper from arXiv. Open the source to read the full record.
arXiv ·
From Qualitative Text to Integrated Evidence: The Parallel-Then-Integrated (PTI) Procedure
Public paper from arXiv. Open the source to read the full record.
arXiv ·
Fantastic Metrics and How to Find Them: Meta-Evaluating Metrics for Multi-lingual Style-Personalized Generation
Public paper from arXiv. Open the source to read the full record.
arXiv ·
HERO: Harnessing Evidence for Test-time Refinement and Prompt Optimization in Program Generation
Public paper from arXiv. Open the source to read the full record.
arXiv ·
How to Evaluate When You Don't Know How to Evaluate? Choosing Reliable Setups for Base Language Model Benchmarking
Public paper from arXiv. Open the source to read the full record.
arXiv ·
NEPSEE: A Double-Annotated Benchmark for Nepali Sentiment Analysis
Public paper from arXiv. Open the source to read the full record.
arXiv ·
Weighted Ensemble Sampling for Rare-Event Estimation in Language Models
Public paper from arXiv. Open the source to read the full record.
arXiv ·
DUALVIEW: Leveraging Broad and Output-Specific Evidence for LLM-based Multi-Output Text Regression
Public paper from arXiv. Open the source to read the full record.
arXiv ·
Titan-Llama: Retrofitting Pre-trained LLMs with External Neural Memory Modules
Public paper from arXiv. Open the source to read the full record.
arXiv ·
Do Syntax Heads Influence Model Behavior?
Public paper from arXiv. Open the source to read the full record.
arXiv ·
Copy-Paste Language Models
Public paper from arXiv. Open the source to read the full record.
arXiv ·
Whose Data, Whose Benchmark? A User-Centric Framework for Auditing Evaluation Resources
Public paper from arXiv. Open the source to read the full record.
arXiv ·
Can LLM Agents Efficiently Identify Appliance Faults? Insights From Diagnosing a Simulated Appliance
Public paper from arXiv. Open the source to read the full record.
arXiv ·
Qwenfant: Train Big Character-level BabyLMs on Concise Parallel Text
Public paper from arXiv. Open the source to read the full record.
arXiv ·
PersEvalBench: Can LLM Evaluators Judge Contextually Appropriate Use of Persona Information?
Public paper from arXiv. Open the source to read the full record.
arXiv ·
TopicMerge: A Divide and Conquer Approach to Topic Modeling
Public paper from arXiv. Open the source to read the full record.
arXiv ·
LLM-jp-4 Technical Report
Public paper from arXiv. Open the source to read the full record.
arXiv ·
Target-Token Asymmetry Can Masquerade As Depth Specialization
Public paper from arXiv. Open the source to read the full record.
arXiv ·
Tail Competition Explains Scaling in Best-of-N Sampling for Verifiable Problems
Public paper from arXiv. Open the source to read the full record.
arXiv ·
Beyond Corrected Actions: Training LLM Agents on Textual Critiques
Public paper from arXiv. Open the source to read the full record.
arXiv ·
Evaluating Compressed Memory: Dependence, Utility, and Content Benefit
Public paper from arXiv. Open the source to read the full record.
Hacker News ·
Step 5 Preview: StepFun's flagship model for agentic work
Hacker News recorded 2 points and 0 comments as of Oct 15, 2026 (UTC).
Hugging Face ·
Welcome Muse Spark Open: open-weight frontier multimodal mixture-of-experts model from Meta
Public article from Hugging Face. Open the source to read the full record.
Hacker News ·
Assessing AI agents at autonomous theoretical physics problem solving
Hacker News recorded 4 points and 0 comments as of Oct 15, 2026 (UTC).
量子位 ·
STEPX首款大模型原生智能体手机STEPX Neo正式发布
Public article from 量子位. Open the source to read the full record.
Google ·
Gemini 4 Argon: Built to solve your hardest problems
We’re giving developers, enterprises, scientists, and consumers access to Gemini 4 Argon, our latest frontier model.
Hacker News ·
OpenStocks, a public earnings forecasting benchmark for AI agents
Hacker News recorded 1 points and 0 comments as of Oct 12, 2026 (UTC).
arXiv ·
AutoIndexer: Flexible-Order Decoding by Training Causal Models on Chains of Edits
Public paper from arXiv. Open the source to read the full record.