SWE-bench reveals that even top language models like Claude 2 resolve only 1.96% of 2,294 real-world GitHub issues, highlighting a gap in practical coding capabilities.
Title resolution pending
16 Pith papers cite this work. Polarity classification is still indexing.
16
Pith papers citing it
representative citing papers
LIMEN discovers effective RL interfaces by using LLMs to evolve observation and reward programs together from raw state, guided by policy training success, outperforming single-component optimization.
DARE reuses up to 87% of attention activations in diffusion LLMs through KV caching and output reuse, delivering 1.2x per-layer latency gains with average performance drops of 1.2-2.0%.
SWE-RL uses RL on software evolution data to train LLMs achieving 41% on SWE-bench Verified with generalization to other reasoning tasks.
RouterBench supplies a standardized benchmark, 405k+ inference dataset, theoretical framework, and comparative analysis for multi-LLM routing systems.
Medusa augments LLMs with multiple decoding heads and tree-based attention to predict and verify several tokens in parallel, yielding 2.2-3.6x inference speedup via two fine-tuning regimes.
Clarification-seeking in LLM agents amplifies prompt injection attack success from ~2% to over 30% across ten frontier models in a new 728-scenario benchmark.
Sync-R1 applies cooperative RL with Sync-GRPO and Dynamic Group Scaling to achieve superior cross-task personalized reasoning in multimodal models on the new UnifyBench++ dataset.
PARSE accelerates LLM inference via parallel semantic prefix verification in a single forward pass, delivering 1.25x-4.3x speedups alone and up to 4.5x when combined with EAGLE-3.
Muon optimizer with weight decay and update scaling achieves ~2x efficiency over AdamW for large LLMs, shown via the Moonlight 3B/16B MoE model trained on 5.7T tokens.
Language models show good calibration when asked to estimate the probability that their own answers are correct, with performance improving as models get larger.
Ranked preference modeling outperforms imitation learning for language model alignment and scales more favorably with model size.
Byte modeling incurs greater scaling overhead for masked diffusion than autoregressive models because the diffusion objective destroys local byte contiguity needed to resolve semantics.
Pruning pretrained MoE models outperforms training from scratch under fixed budget, different expert compression methods converge after continued training, and progressive pruning plus multi-token KD improves the final 23A2B model.
DACA-GRPO adds denoising-aware credit assignment and bias-reduced likelihood estimation to GRPO, delivering consistent gains up to 36.3pp on math, code, constraint, and schema benchmarks for diffusion LLMs.
citing papers explorer
-
SWE-bench: Can Language Models Resolve Real-World GitHub Issues?
SWE-bench reveals that even top language models like Claude 2 resolve only 1.96% of 2,294 real-world GitHub issues, highlighting a gap in practical coding capabilities.
-
Discovering Reinforcement Learning Interfaces with Large Language Models
LIMEN discovers effective RL interfaces by using LLMs to evolve observation and reward programs together from raw state, guided by policy training success, outperforming single-component optimization.
-
DARE: Diffusion Language Model Activation Reuse for Efficient Inference
DARE reuses up to 87% of attention activations in diffusion LLMs through KV caching and output reuse, delivering 1.2x per-layer latency gains with average performance drops of 1.2-2.0%.
-
SWE-RL: Advancing LLM Reasoning via Reinforcement Learning on Open Software Evolution
SWE-RL uses RL on software evolution data to train LLMs achieving 41% on SWE-bench Verified with generalization to other reasoning tasks.
-
RouterBench: A Benchmark for Multi-LLM Routing System
RouterBench supplies a standardized benchmark, 405k+ inference dataset, theoretical framework, and comparative analysis for multi-LLM routing systems.
-
Medusa: Simple LLM Inference Acceleration Framework with Multiple Decoding Heads
Medusa augments LLMs with multiple decoding heads and tree-based attention to predict and verify several tokens in parallel, yielding 2.2-3.6x inference speedup via two fine-tuning regimes.
-
ASPI: Seeking Ambiguity Clarification Amplifies Prompt Injection Vulnerability in LLM Agents
Clarification-seeking in LLM agents amplifies prompt injection attack success from ~2% to over 30% across ten frontier models in a new 728-scenario benchmark.
-
Uni-Synergy: Bridging Understanding and Generation for Personalized Reasoning via Co-operative Reinforcement Learning
Sync-R1 applies cooperative RL with Sync-GRPO and Dynamic Group Scaling to achieve superior cross-task personalized reasoning in multimodal models on the new UnifyBench++ dataset.
-
Parallel Prefix Verification for Speculative Generation
PARSE accelerates LLM inference via parallel semantic prefix verification in a single forward pass, delivering 1.25x-4.3x speedups alone and up to 4.5x when combined with EAGLE-3.
-
Muon is Scalable for LLM Training
Muon optimizer with weight decay and update scaling achieves ~2x efficiency over AdamW for large LLMs, shown via the Moonlight 3B/16B MoE model trained on 5.7T tokens.
-
Language Models (Mostly) Know What They Know
Language models show good calibration when asked to estimate the probability that their own answers are correct, with performance improving as models get larger.
-
A General Language Assistant as a Laboratory for Alignment
Ranked preference modeling outperforms imitation learning for language model alignment and scales more favorably with model size.
-
The Efficiency Gap in Byte Modeling
Byte modeling incurs greater scaling overhead for masked diffusion than autoregressive models because the diffusion objective destroys local byte contiguity needed to resolve semantics.
-
SlimQwen: Exploring the Pruning and Distillation in Large MoE Model Pre-training
Pruning pretrained MoE models outperforms training from scratch under fixed budget, different expert compression methods converge after continued training, and progressive pruning plus multi-token KD improves the final 23A2B model.
-
DACA-GRPO: Denoising-Aware Credit Assignment for Reinforcement Learning in Diffusion Language Models
DACA-GRPO adds denoising-aware credit assignment and bias-reduced likelihood estimation to GRPO, delivering consistent gains up to 36.3pp on math, code, constraint, and schema benchmarks for diffusion LLMs.
- Learning in the Fisher Subspace: A Guided Initialization for LoRA Fine-Tuning