LazyAttention kernelizes deferred positional encoding to enable zero-copy, position-agnostic KV cache reuse, delivering 1.37× lower TTFT and 1.40× higher throughput than Block-Attention under skewed document distributions while preserving output quality.
hub
The N arrative QA reading comprehension challenge
16 Pith papers cite this work. Polarity classification is still indexing.
16
Pith papers citing it
hub tools
citation-role summary
dataset 1
citation-polarity summary
roles
dataset 1polarities
use dataset 1representative citing papers
R^2-Mem distills rubric-scored experiences from high- and low-quality search trajectories to guide LLM agents, raising F1 by up to 22.6% while cutting tokens 12.9% and iterations 20.2%.
LLM novel summaries emphasize endings more than human ones, measured by aligning summary sentences to referenced chapters.
Develops an information-theoretic framework showing surprise and coherence trade off in single reader models but coexist via pre- and post-revelation modes, operationalized as reference-less LLM metrics for fair play and validated on generated stories plus classic detective fiction.
M3-Embedding is a single model for multi-lingual, multi-functional, and multi-granular text embeddings trained via self-knowledge distillation that achieves new state-of-the-art results on multilingual, cross-lingual, and long-document retrieval benchmarks.
GAIA benchmark shows humans at 92% accuracy on simple real-world questions far outperform current AI systems at 15%, proposing this gap as a key milestone for general AI.
BIG-bench is a 204-task benchmark that measures scaling trends, calibration, and absolute limitations of language models across knowledge, reasoning, and social domains.
SKOP uses key-orthogonal projections to steer LLM activations while preserving attention patterns on focus tokens, cutting utility degradation by 5-7x and retaining over 95% of standard steering efficacy.
Mixture-of-experts flow matching enables non-autoregressive language models to achieve autoregressive-level quality in three sampling steps, delivering up to 1000x faster inference than diffusion models.
ForkKV uses copy-on-write disaggregated KV cache with DualRadixTree and ResidualAttention kernels to deliver up to 3x throughput over prior multi-LoRA serving systems with negligible quality loss.
LLMs achieve higher accuracy than humans on compositional imagery tasks previously argued to require pictorial representations, supporting emergent propositional mental imagery in AI.
LIMO achieves 63.3% on AIME24 and 95.6% on MATH500 via supervised fine-tuning on roughly 1% of the data used by prior models, supporting the claim that minimal strategic examples suffice when pre-training has already encoded domain knowledge.
Quest speeds up long-context LLM self-attention by up to 2.23x via query-dependent selection of top-K critical KV cache pages, cutting overall latency by 7.03x with negligible accuracy loss.
Gemini Ultra reaches human-expert performance on MMLU for the first time and sets new state-of-the-art results on 30 of 32 benchmarks, including all 20 multimodal ones tested.
GRKV applies global ridge regression to KV cache merging for span-based retention in long-context LLMs, claiming to be the only method that improves benchmark performance with minimal overhead.
Gated DeltaNet integrates gating and delta rules into linear transformers, outperforming Mamba2 and DeltaNet on language modeling, reasoning, retrieval, and long-context tasks.
citing papers explorer
-
LazyAttention: Efficient Retrieval-Augmented Generation with Deferred Positional Encoding
LazyAttention kernelizes deferred positional encoding to enable zero-copy, position-agnostic KV cache reuse, delivering 1.37× lower TTFT and 1.40× higher throughput than Block-Attention under skewed document distributions while preserving output quality.
-
R^2-Mem: Reflective Experience for Memory Search
R^2-Mem distills rubric-scored experiences from high- and low-quality search trajectories to guide LLM agents, raising F1 by up to 22.6% while cutting tokens 12.9% and iterations 20.2%.
-
Attention Flows: Tracing LLM Conceptual Engagement via Story Summaries
LLM novel summaries emphasize endings more than human ones, measured by aligning summary sentences to referenced chapters.
-
The Challenge and Reward of Fair Play in Narrative: A Computational Approach
Develops an information-theoretic framework showing surprise and coherence trade off in single reader models but coexist via pre- and post-revelation modes, operationalized as reference-less LLM metrics for fair play and validated on generated stories plus classic detective fiction.
-
M3-Embedding: Multi-Linguality, Multi-Functionality, Multi-Granularity Text Embeddings Through Self-Knowledge Distillation
M3-Embedding is a single model for multi-lingual, multi-functional, and multi-granular text embeddings trained via self-knowledge distillation that achieves new state-of-the-art results on multilingual, cross-lingual, and long-document retrieval benchmarks.
-
GAIA: a benchmark for General AI Assistants
GAIA benchmark shows humans at 92% accuracy on simple real-world questions far outperform current AI systems at 15%, proposing this gap as a key milestone for general AI.
-
Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language models
BIG-bench is a 204-task benchmark that measures scaling trends, calibration, and absolute limitations of language models across knowledge, reasoning, and social domains.
-
Don't Lose Focus: Activation Steering via Key-Orthogonal Projections
SKOP uses key-orthogonal projections to steer LLM activations while preserving attention patterns on focus tokens, cutting utility degradation by 5-7x and retaining over 95% of standard steering efficacy.
-
Towards Faster Language Model Inference Using Mixture-of-Experts Flow Matching
Mixture-of-experts flow matching enables non-autoregressive language models to achieve autoregressive-level quality in three sampling steps, delivering up to 1000x faster inference than diffusion models.
-
ForkKV: Scaling Multi-LoRA Agent Serving via Copy-on-Write Disaggregated KV Cache
ForkKV uses copy-on-write disaggregated KV cache with DualRadixTree and ResidualAttention kernels to deliver up to 3x throughput over prior multi-LoRA serving systems with negligible quality loss.
-
Artificial Phantasia: Emergent Mental Imagery in Large Language Models
LLMs achieve higher accuracy than humans on compositional imagery tasks previously argued to require pictorial representations, supporting emergent propositional mental imagery in AI.
-
LIMO: Less is More for Reasoning
LIMO achieves 63.3% on AIME24 and 95.6% on MATH500 via supervised fine-tuning on roughly 1% of the data used by prior models, supporting the claim that minimal strategic examples suffice when pre-training has already encoded domain knowledge.
-
Quest: Query-Aware Sparsity for Efficient Long-Context LLM Inference
Quest speeds up long-context LLM self-attention by up to 2.23x via query-dependent selection of top-K critical KV cache pages, cutting overall latency by 7.03x with negligible accuracy loss.
-
Gemini: A Family of Highly Capable Multimodal Models
Gemini Ultra reaches human-expert performance on MMLU for the first time and sets new state-of-the-art results on 30 of 32 benchmarks, including all 20 multimodal ones tested.
-
GRKV: Global Regression for Training-Free KV Cache Compression in Long-Context LLMs
GRKV applies global ridge regression to KV cache merging for span-based retention in long-context LLMs, claiming to be the only method that improves benchmark performance with minimal overhead.
-
Gated Delta Networks: Improving Mamba2 with Delta Rule
Gated DeltaNet integrates gating and delta rules into linear transformers, outperforming Mamba2 and DeltaNet on language modeling, reasoning, retrieval, and long-context tasks.