inference optimization

An Empirical Study of LLM Reasoning Ability Under Strict Output Length Constraint EMNLP 2025 Answer Convergence as a Signal for Early Stopping in Reasoning EMNLP 2025 A Training-Free Sub-quadratic Cost Transformer Model Serving Framework with Hierarchically Pruned Attention ICLR 2025 A-VL: Adaptive Attention for Large Vision-Language Models AAAI 2025 BigMac: A Communication-Efficient Mixture-of-Experts Model Structure for Fast Training and Inference AAAI 2025 Brevity is the soul of sustainability: Characterizing LLM response lengths ACL 2025 C3oT: Generating Shorter Chain-of-Thought Without Compromising Effectiveness AAAI 2025 Cacheback: Speculative Decoding With Nothing But Cache EMNLP 2025 CacheQuant: Comprehensively Accelerated Diffusion Models CVPR 2025 CLaSp: In-Context Layer Skip for Self-Speculative Decoding ACL 2025 ClusterAttn: KV Cache Compression under Intrinsic Attention Clustering ACL 2025 Confidence-Aware Reasoning: Optimizing Self-Guided Thinking Trajectories in Large Reasoning Models EMNLP 2025 CoTFormer: A Chain of Thought Driven Architecture with Budget-Adaptive Computation Cost at Inference ICLR 2025 DAST: Difficulty-Adaptive Slow-Thinking for Large Reasoning Models EMNLP 2025 Data-Free Group-Wise Fully Quantized Winograd Convolution via Learnable Scales CVPR 2025 Decoding Speculative Decoding NAACL 2025 DeFT: Decoding with Flash Tree-attention for Efficient Tree-structured LLM Inference ICLR 2025 DiffSkip: Differential Layer Skipping in Large Language Models ACL 2025 Distributed Speculative Inference (DSI): Speculation Parallelism for Provably Faster Lossless Language Model Inference ICLR 2025 DivPrune: Diversity-based Visual Token Pruning for Large Multimodal Models CVPR 2025 Does quantization affect models’ performance on long-context tasks? EMNLP 2025 Don’t Think It Twice: Exploit Shift Invariance for Efficient Online Streaming Inference of CNNs AAAI 2025 Draft Model Knows When to Stop: Self-Verification Speculative Decoding for Long-Form Generation EMNLP 2025 DS-LLM: Leveraging Dynamical Systems to Enhance Both Training and Inference of Large Language Models ICLR 2025 Dynamic Parallel Tree Search for Efficient LLM Reasoning ACL 2025