4 papers · 1 filter
Code as Agent Harness
Xuying Ning, Katherine Tieu, Dongqi Fu +39
Recent large language models (LLMs) have demonstrated strong capabilities in understanding and generating code, from competitive programming to repository-level software engineerin…
AdaFuse: Adaptive Ensemble Decoding with Test-Time Scaling for LLMs
Chengming Cui, Tianxin Wei, Ziyi Chen +6
Large language models (LLMs) exhibit complementary strengths arising from differences in pretraining data, model architectures, and decoding behaviors. Inference-time ensembling pr…
Haystack Engineering: Context Engineering for Heterogeneous and Agentic Long-Context Evaluation
Mufei Li, Dongqi Fu, Limei Wang +10
Modern long-context large language models (LLMs) perform well on synthetic "needle-in-a-haystack" (NIAH) benchmarks, but such tests overlook how noisy contexts arise from biased re…
Transformer Copilot: Learning from The Mistake Log in LLM Fine-tuning
Jiaru Zou, Yikun Ban, Zihao Li +4
Large language models are typically adapted to downstream tasks through supervised fine-tuning on domain-specific data. While standard fine-tuning focuses on minimizing generation…