Sieve dynamically schedules MoE experts across GPU and PIM hardware to handle bimodal token distributions, achieving 1.3x to 1.6x gains in throughput and interactivity over static prior PIM systems on three large models.
super hub
Mixtral of Experts
125 Pith papers cite this work. Polarity classification is still indexing.
125
Pith papers citing it
abstract
We introduce Mixtral 8x7B, a Sparse Mixture of Experts (SMoE) language model. Mixtral has the same architecture as Mistral 7B, with the difference that each layer is composed of 8 feedforward blocks (i.e. experts). For every token, at each layer, a router network selects two experts to process the current state and combine their outputs. Even though each token only sees two experts, the selected experts can be different at each timestep. As a result, each token has access to 47B parameters, but only uses 13B active parameters during inference. Mixtral was trained with a context size of 32k tokens and it outperforms or matches Llama 2 70B and GPT-3.5 across all evaluated benchmarks. In particular, Mixtral vastly outperforms Llama 2 70B on mathematics, code generation, and multilingual benchmarks. We also provide a model fine-tuned to follow instructions, Mixtral 8x7B - Instruct, that surpasses GPT-3.5 Turbo, Claude-2.1, Gemini Pro, and Llama 2 70B - chat model on human benchmarks. Both the base and instruct models are released under the Apache 2.0 license.
hub tools
citation-role summary
background 2
citation-polarity summary
claims ledger
- abstract We introduce Mixtral 8x7B, a Sparse Mixture of Experts (SMoE) language model. Mixtral has the same architecture as Mistral 7B, with the difference that each layer is composed of 8 feedforward blocks (i.e. experts). For every token, at each layer, a router network selects two experts to process the current state and combine their outputs. Even though each token only sees two experts, the selected experts can be different at each timestep. As a result, each token has access to 47B parameters, but only uses 13B active parameters during inference. Mixtral was trained with a context size of 32k tok
authors
co-cited works
roles
background 2polarities
background 2representative citing papers
ReLibra uses pre-known token-to-expert routing from RL rollouts to perform inter-batch expert reordering and intra-batch replication, delivering up to 1.6x higher throughput than Megatron-LM and 1.2x over oracle-equipped EPLB while staying within 6-10% of an ideal balanced baseline.
RoundPipe achieves near-zero-bubble pipeline parallelism for LLM training on consumer GPUs by dynamically dispatching computation stages round-robin, yielding 1.48-2.16x speedups and enabling 235B model fine-tuning on 8x RTX 4090.
The AI Scientist framework enables LLMs to independently conduct the full scientific process from idea generation to paper writing and review, demonstrated across three ML subfields with papers costing under $15 each.
OSWorld provides the first unified real-computer benchmark for open-ended multimodal agent tasks, exposing large performance gaps between humans and state-of-the-art LLM/VLM agents.
RULER shows most long-context LMs drop sharply in performance on complex tasks as length and difficulty increase, with only half maintaining results at 32K tokens.
Routers in SMoE models form geometric alignments with their experts through shared gradient directions, enabling effective specialization that auxiliary load-balancing losses tend to disrupt.
Hebatron is the first open-weight Hebrew MoE LLM adapted from Nemotron-3, reaching 73.8% on Hebrew reasoning benchmarks while activating only 3B parameters per pass and supporting 65k-token context.
Dooly reduces LLM inference profiling costs by 56.4% via configuration-agnostic taint-based labeling and selective database reuse, delivering simulation accuracy within 5% MAPE for TTFT and 8% for TPOT across 12 models.
Evaluation artifacts substantially inflate the measured unsolvability ceiling in multi-LLM routing, leading to distorted router training and overstated headroom.
MISA routes to a small subset of indexer heads via block statistics, matching full DSA performance on LongBench with 4-8x fewer heads and 3.82x speedup while recovering over 92% of selected tokens.
Standard top-k routers in MoE language models often select suboptimal routes for difficult tokens, and updating only the final router layer raises pass@K on AIME and HMMT benchmarks across multiple models.
Every fixed finite feedforward neural network definable in an o-minimal structure has finite sample complexity in the agnostic PAC setting.
BatMIL uses hybrid hyperbolic-Euclidean geometry, an S4 state-space backbone, and chunk-level mixture-of-experts to outperform prior multiple-instance learning methods on seven whole-slide image datasets across six cancers.
Coral cuts multi-LLM serving costs by up to 2.79x and raises goodput by up to 2.39x on heterogeneous GPUs through adaptive joint optimization and a lossless two-stage decomposition that solves quickly.
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%.
Adding temporal memory via LIF, precision-weighted gating, and anticipatory prediction to MoE routers recovers effective expert selection at distribution transitions, with ablation confirming a super-additive beta-ant interaction.
MASCing uses an LSTM surrogate and optimized steering masks to enable flexible, inference-time control over MoE expert routing for safety objectives, improving jailbreak defense and content generation success rates substantially across multiple models.
Machine collective intelligence uses coordinated AI agents to evolve symbolic hypotheses and recover governing equations from observations in deterministic, stochastic, and uncharacterized systems, achieving up to six orders of magnitude better extrapolation than neural networks with 5-40 parameters
ACO-MoE recovers 95.3% of clean-input performance in visual control tasks under Markov-switching corruptions by routing restoration experts and anchoring representations to clean foreground masks.
Incompressible Knowledge Probes enable log-linear estimation of LLM parameter counts from factual accuracy on obscure questions, showing continued scaling of knowledge capacity across open and closed models.
Bayesian softmax-gated mixture-of-experts models achieve posterior contraction for density estimation and parameter recovery using Voronoi losses, plus two strategies for choosing the number of experts.
Expert upcycling duplicates experts in an existing MoE checkpoint and continues pre-training to match fixed-size baseline performance with 32% less compute.
GEM replaces learned routers in MoE models with a global scheduler based on linear programming relaxation and hierarchical rounding, achieving SOTA on the UODB multi-domain benchmark with gains on rare domains.
citing papers explorer
-
Sieve: Dynamic Expert-Aware PIM Acceleration for Evolving Mixture-of-Experts Models
Sieve dynamically schedules MoE experts across GPU and PIM hardware to handle bimodal token distributions, achieving 1.3x to 1.6x gains in throughput and interactivity over static prior PIM systems on three large models.
-
ReLibra: Routing-Replay-Guided Load Balancing for MoE Training in Reinforcement Learning
ReLibra uses pre-known token-to-expert routing from RL rollouts to perform inter-batch expert reordering and intra-batch replication, delivering up to 1.6x higher throughput than Megatron-LM and 1.2x over oracle-equipped EPLB while staying within 6-10% of an ideal balanced baseline.
-
Efficient Training on Multiple Consumer GPUs with RoundPipe
RoundPipe achieves near-zero-bubble pipeline parallelism for LLM training on consumer GPUs by dynamically dispatching computation stages round-robin, yielding 1.48-2.16x speedups and enabling 235B model fine-tuning on 8x RTX 4090.
-
The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery
The AI Scientist framework enables LLMs to independently conduct the full scientific process from idea generation to paper writing and review, demonstrated across three ML subfields with papers costing under $15 each.
-
OSWorld: Benchmarking Multimodal Agents for Open-Ended Tasks in Real Computer Environments
OSWorld provides the first unified real-computer benchmark for open-ended multimodal agent tasks, exposing large performance gaps between humans and state-of-the-art LLM/VLM agents.
-
RULER: What's the Real Context Size of Your Long-Context Language Models?
RULER shows most long-context LMs drop sharply in performance on complex tasks as length and difficulty increase, with only half maintaining results at 32K tokens.
-
Routers Learn the Geometry of Their Experts: Geometric Coupling in Sparse Mixture-of-Experts
Routers in SMoE models form geometric alignments with their experts through shared gradient directions, enabling effective specialization that auxiliary load-balancing losses tend to disrupt.
-
HEBATRON: A Hebrew-Specialized Open-Weight Mixture-of-Experts Language Model
Hebatron is the first open-weight Hebrew MoE LLM adapted from Nemotron-3, reaching 73.8% on Hebrew reasoning benchmarks while activating only 3B parameters per pass and supporting 65k-token context.
-
Dooly: Configuration-Agnostic, Redundancy-Aware Profiling for LLM Inference Simulation
Dooly reduces LLM inference profiling costs by 56.4% via configuration-agnostic taint-based labeling and selective database reuse, delivering simulation accuracy within 5% MAPE for TTFT and 8% for TPOT across 12 models.
-
Unsolvability Ceiling in Multi-LLM Routing: An Empirical Study of Evaluation Artifacts
Evaluation artifacts substantially inflate the measured unsolvability ceiling in multi-LLM routing, leading to distorted router training and overstated headroom.
-
MISA: Mixture of Indexer Sparse Attention for Long-Context LLM Inference
MISA routes to a small subset of indexer heads via block statistics, matching full DSA performance on LongBench with 4-8x fewer heads and 3.82x speedup while recovering over 92% of selected tokens.
-
When Are Experts Misrouted? Counterfactual Routing Analysis in Mixture-of-Experts Language Models
Standard top-k routers in MoE language models often select suboptimal routes for difficult tokens, and updating only the final router layer raises pass@K on AIME and HMMT benchmarks across multiple models.
-
Every Feedforward Neural Network Definable in an o-Minimal Structure Has Finite Sample Complexity
Every fixed finite feedforward neural network definable in an o-minimal structure has finite sample complexity in the agnostic PAC setting.
-
Geometry-Aware State Space Model: A New Paradigm for Whole-Slide Image Representation
BatMIL uses hybrid hyperbolic-Euclidean geometry, an S4 state-space backbone, and chunk-level mixture-of-experts to outperform prior multiple-instance learning methods on seven whole-slide image datasets across six cancers.
-
Coral: Cost-Efficient Multi-LLM Serving over Heterogeneous Cloud GPUs
Coral cuts multi-LLM serving costs by up to 2.79x and raises goodput by up to 2.39x on heterogeneous GPUs through adaptive joint optimization and a lossless two-stage decomposition that solves quickly.
-
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%.
-
Affinity Is Not Enough: Recovering the Free Energy Principle in Mixture-of-Experts
Adding temporal memory via LIF, precision-weighted gating, and anticipatory prediction to MoE routers recovers effective expert selection at distribution transitions, with ablation confirming a super-additive beta-ant interaction.
-
MASCing: Configurable Mixture-of-Experts Behavior via Activation Steering Masks
MASCing uses an LSTM surrogate and optimized steering masks to enable flexible, inference-time control over MoE expert routing for safety objectives, improving jailbreak defense and content generation success rates substantially across multiple models.
-
Machine Collective Intelligence for Explainable Scientific Discovery
Machine collective intelligence uses coordinated AI agents to evolve symbolic hypotheses and recover governing equations from observations in deterministic, stochastic, and uncharacterized systems, achieving up to six orders of magnitude better extrapolation than neural networks with 5-40 parameters
-
Agent-Centric Observation Adaptation for Robust Visual Control under Dynamic Perturbations
ACO-MoE recovers 95.3% of clean-input performance in visual control tasks under Markov-switching corruptions by routing restoration experts and anchoring representations to clean foreground masks.
-
Incompressible Knowledge Probes: Estimating Black-Box LLM Parameter Counts via Factual Capacity
Incompressible Knowledge Probes enable log-linear estimation of LLM parameter counts from factual accuracy on obscure questions, showing continued scaling of knowledge capacity across open and closed models.
-
On Bayesian Softmax-Gated Mixture-of-Experts Models
Bayesian softmax-gated mixture-of-experts models achieve posterior contraction for density estimation and parameter recovery using Voronoi losses, plus two strategies for choosing the number of experts.
-
Expert Upcycling: Shifting the Compute-Efficient Frontier of Mixture-of-Experts
Expert upcycling duplicates experts in an existing MoE checkpoint and continues pre-training to match fixed-size baseline performance with 32% less compute.
-
Multi-Domain Learning with Global Expert Mapping
GEM replaces learned routers in MoE models with a global scheduler based on linear programming relaxation and hierarchical rounding, achieving SOTA on the UODB multi-domain benchmark with gains on rare domains.
-
GTA-2: Benchmarking General Tool Agents from Atomic Tool-Use to Open-Ended Workflows
GTA-2 benchmark shows frontier models achieve below 50% on atomic tool tasks and only 14.39% success on realistic long-horizon workflows, with execution harnesses like Manus providing substantial gains.
-
SecureRouter: Encrypted Routing for Efficient Secure Inference
SecureRouter accelerates secure transformer inference by 1.95x via an encrypted router that selects input-adaptive models from an MPC-optimized pool with negligible accuracy loss.
-
Atropos: Improving Cost-Benefit Trade-off of LLM-based Agents under Self-Consistency with Early Termination and Model Hotswap
Atropos uses GCN on inference graphs for early failure prediction and hotswaps to larger LLMs, achieving 74% of large-model performance at 24% cost.
-
Equifinality in Mixture of Experts: Routing Topology Does Not Determine Language Modeling Quality
Routing topology in sparse Mixture-of-Experts models does not determine asymptotic language modeling perplexity; multiple variants including cosine-similarity routing achieve statistically equivalent performance.
-
A Sanity Check on Composed Image Retrieval
The paper creates FISD, a controlled benchmark for composed image retrieval that removes query ambiguity via generative models, and proposes a multi-round agentic evaluation to assess models in interactive settings.
-
A Full-Stack Performance Evaluation Infrastructure for 3D-DRAM-based LLM Accelerators
ATLAS is the first silicon-validated simulation framework for 3D-DRAM LLM accelerators, achieving under 8.57% error and over 97% correlation with real hardware while supporting design exploration.
-
An Agentic Evaluation Architecture for Historical Bias Detection in Educational Textbooks
An agentic architecture with multimodal screening, a five-agent jury, meta-synthesis, and source attribution protocol detects biases in Romanian history textbooks more accurately than zero-shot baselines, achieving 83.3% acceptable excerpts and human preference in 64.8% of blind comparisons.
-
InfiniLoRA: Disaggregated Multi-LoRA Serving for Large Language Models
InfiniLoRA decouples LoRA execution from base-model inference and reports 3.05x higher request throughput plus 54% more adapters meeting strict latency SLOs.
-
A Switch-Centric In-Network Architecture for Accelerating LLM Inference in Shared-Memory Network
SCIN uses an in-switch accelerator for direct memory access and 8-bit in-network quantization during All-Reduce, delivering up to 8.7x faster small-message reduction and 1.74x TTFT speedup on LLaMA-2 models.
-
Agent Security Bench (ASB): Formalizing and Benchmarking Attacks and Defenses in LLM-based Agents
ASB is a new benchmark that tests 10 prompt injection attacks, memory poisoning, a novel Plan-of-Thought backdoor attack, and 11 defenses on LLM agents across 13 models, finding attack success rates up to 84.3% and limited defense effectiveness.
-
Jamba: A Hybrid Transformer-Mamba Language Model
Jamba presents a hybrid Transformer-Mamba MoE architecture for LLMs that delivers state-of-the-art benchmark performance and strong results up to 256K token contexts while fitting in one 80GB GPU with high throughput.
-
N-vium: Mixture-of-Exits Transformer for Accelerated Exact Generation
N-vium achieves 57.9% wall-clock speedup over matched standard transformers at no perplexity cost by mixing exact predictions from multiple model depths.
-
When Attention Closes: How LLMs Lose the Thread in Multi-Turn Interaction
Attention to goal tokens declines in multi-turn LLM interactions while residual representations often retain decodable goal information, and the gap between these predicts whether goal-conditioned behavior survives.
-
Instruction Lens Score: Your Instruction Contributes a Powerful Object Hallucination Detector for Multimodal Large Language Models
Instruction token embeddings encode visual information that can be leveraged to detect object hallucinations in MLLMs via a new combined score outperforming prior detectors.
-
A Unified Graph Language Model for Multi-Domain Multi-Task Graph Alignment Instruction Tuning
UniGraphLM uses a multi-domain multi-task GNN encoder and adaptive alignment to create unified graph tokens for LLMs across diverse domains and tasks.
-
ROMER: Expert Replacement and Router Calibration for Robust MoE LLMs on Analog Compute-in-Memory Systems
ROMER cuts perplexity by up to 59% in noisy analog CIM environments for MoE LLMs via expert replacement and router recalibration calibrated on real-chip measurements.
-
MLCommons Chakra: Advancing Performance Benchmarking and Co-design using Standardized Execution Traces
Chakra introduces a portable, interoperable graph-based execution trace format for distributed ML workloads along with supporting tools to standardize performance benchmarking and software-hardware co-design.
-
Instructions Shape Production of Language, not Processing
Instructions trigger a production-centered mechanism in language models, with task-specific information stable in input tokens but varying strongly in output tokens and correlating with behavior.
-
Exploring Token-Space Manipulation in Latent Audio Tokenizers
LATTE creates a compact latent token bottleneck in audio tokenizers that aggregates global information and enables unsupervised editing of attributes like speaker identity via token swapping.
-
Enabling Performant and Flexible Model-Internal Observability for LLM Inference
DMI-Lib delivers 0.4-6.8% overhead for offline batch LLM inference and ~6% for moderate online serving while exposing rich internal signals across backends, cutting latency overhead 2-15x versus prior observability baselines.
-
DECO: Sparse Mixture-of-Experts with Dense-Comparable Performance on End-Side Devices
DECO matches dense model performance at 20% expert activation via ReLU-based routing with learnable scaling and the NormSiLU activation, plus a 3x real-hardware speedup.
-
NanoResearch: Co-Evolving Skills, Memory, and Policy for Personalized Research Automation
NanoResearch introduces a tri-level co-evolving framework of skills, memory, and policy to personalize LLM-powered research automation across projects and users.
-
LegalCiteBench: Evaluating Citation Reliability in Legal Language Models
LegalCiteBench reveals that current LLMs achieve under 7% accuracy on closed-book legal citation retrieval and completion tasks, with misleading answer rates above 94% for nearly all tested models.
-
DisagMoE: Computation-Communication overlapped MoE Training via Disaggregated AF-Pipe Parallelism
DisagMoE achieves up to 1.8x faster MoE training by disaggregating attention and FFN layers into disjoint GPU groups with a multi-stage uni-directional pipeline and roofline-based bandwidth balancing.
-
Sparse Layers are Critical to Scaling Looped Language Models
Looped MoE models scale better than standard transformers because different experts activate on each loop pass, recovering expressivity without extra parameters, and support superior early exits.
-
SlimQwen: Exploring the Pruning and Distillation in Large MoE Model Pre-training
Pruning pretrained MoE models outperforms training from scratch, different compression methods converge after continued pretraining, and combining KD with language modeling loss plus progressive schedules yields a competitive 23A2B model from Qwen3-Next-80A3B.