NeurIPS 2026 reviews are up: reviews, scores, rebuttals and meta-reviews for 7,877 papers. See the highest rated
acceptodds: the papers of ICLR 2027
180 results for “keyword:"llm inference"” in ICLR 2027. 24 withdrawn or desk rejected listed apart. clearsort:relevanceratingnewestmost read
[PDF]From Chains to Trees: Parent-Conditioned Drafting for Semi-Autoregressive Speculative Decoding▼4.3k2.00
[PDF]A Memory You Can’t Recall Is No Memory at All: MEMENTO for Query-Agnostic KV Cache Eviction in Multi-Turn Inference▼5.8k2.33
[PDF]Fail Fast, Win Big: Rethinking the Drafting Strategy in Speculative Decoding via Diffusion LLMs▼4.3k2.00
[PDF]HeteroLoRA-MoE: Heterogeneous Low-Rank Experts with Diverse Projection Patterns and Efficient GPU Kernels▼4.3k2.00
[PDF]Compression-Aware Fine-Tuning: Adapting Language Model Weights to a Frozen Cache Bottleneck▼4.3k2.00
[PDF]Focus Calibration for Ultra-Low-Bit All-Reduce in Bandwidth-Constrained Collaborative LLM Inference▼4.3k2.00
[PDF]SpecSplice: Dynamic Composition of Parametric and Retrieval-based Drafts for Agentic Speculative Decoding▼1.2k3.00
[PDF]UNDERSTANDING MOE INFERENCE ENERGY CONSUMPTION VIA EXPERT ACTIVATION TRAJECTORY ANALYSIS▼1361.33
[PDF]From Rigid to Dynamic: Entropy-Guided Adaptive Inference for Accelerating Long-Context LLMs▼1.2k1.67
[PDF]Reciprocal Guidance: Orchestrating Draft and Verify Budgets for Advancing the Diffusion–AR Self-Speculation Frontier▼5.8k2.33
[PDF]DSpark-OPD: Acceptance-Aligned On-Policy Distillation for Semi-Autoregressive Speculative Drafting▼1.2k1.67
[PDF]QSpar: Co-Designing Sparse Attention and Quantized Weights for Long-Context Self-Speculative Decoding▲5103.00
Withdrawn or desk rejected (24)
[PDF]DyS-MoE: Virtualizing Dynamic Expert Execution with Static Windows for Resource-Constrained MoE InferenceWithdrawn2.00
[PDF]SpecRelay: Amortize or Avoid the Round Trip in Edge–Cloud Speculative DecodingDesk rejected2.00
[PDF]Easy-Kuant: Diagnosing and Mitigating KV-Cache Quantization Collapse with Selective Layer ProtectionDesk rejected