UltraEP is the first exact-load real-time expert balancer for large-EP MoE training and serving on rack-scale nodes, reaching 94.3% of ideal throughput and 1.49x over no-balancing.
super hub Mixed citations
GLM-4.5: Agentic, Reasoning, and Coding (ARC) Foundation Models
Mixed citation behavior. Most common role is background (69%).
130
Pith papers citing it
Background
69% of classified citations
abstract
We present GLM-4.5, an open-source Mixture-of-Experts (MoE) large language model with 355B total parameters and 32B activated parameters, featuring a hybrid reasoning method that supports both thinking and direct response modes. Through multi-stage training on 23T tokens and comprehensive post-training with expert model iteration and reinforcement learning, GLM-4.5 achieves strong performance across agentic, reasoning, and coding (ARC) tasks, scoring 70.1% on TAU-Bench, 91.0% on AIME 24, and 64.2% on SWE-bench Verified. With much fewer parameters than several competitors, GLM-4.5 ranks 3rd overall among all evaluated models and 2nd on agentic benchmarks. We release both GLM-4.5 (355B parameters) and a compact version, GLM-4.5-Air (106B parameters), to advance research in reasoning and agentic AI systems. Code, models, and more information are available at https://github.com/zai-org/GLM-4.5.
hub tools
citation-role summary
background 19
baseline 3
method 3
other 1
citation-polarity summary
claims ledger
- abstract We present GLM-4.5, an open-source Mixture-of-Experts (MoE) large language model with 355B total parameters and 32B activated parameters, featuring a hybrid reasoning method that supports both thinking and direct response modes. Through multi-stage training on 23T tokens and comprehensive post-training with expert model iteration and reinforcement learning, GLM-4.5 achieves strong performance across agentic, reasoning, and coding (ARC) tasks, scoring 70.1% on TAU-Bench, 91.0% on AIME 24, and 64.2% on SWE-bench Verified. With much fewer parameters than several competitors, GLM-4.5 ranks 3rd ove
authors
co-cited works
representative citing papers
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 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.
WildTableBench is the first QA benchmark for naturally occurring table images, where 21 multimodal models were evaluated and only one exceeded 50% accuracy.
User-turn generation reveals that LLMs' interaction awareness is largely decoupled from task accuracy, remaining near zero in deterministic settings even as accuracy scales to 96.8% on GSM8K.
Static SFT and RL training for tool-use agents leads to performance drops under open-world distributional shifts across perception, interaction, reasoning and internalization; perturbation-augmented fine-tuning is proposed as mitigation.
Proposes Monotonic Inference Policy Improvement (MIPI) objective and MIPU two-step update framework to address objective misalignment between training and inference policies in LLM reinforcement learning.
MacroLens is a point-in-time multi-signal benchmark dataset and seven tasks for evaluating contextual financial reasoning models under macroeconomic scenarios.
Agent-generated code produces up to 13.1% lower follow-on task resolution rates than human code in chained repository-level experiments, with differences linked to behavioral patterns rather than conventional metrics.
Empirical evaluation on LiveCodeBench shows certainty-based RLIF yields early gains followed by output shortening and reasoning collapse, providing no advantage for RLVR initialization on code tasks.
LegalWorld is a life-cycle interactive environment modeling Chinese civil litigation as five causally connected stages grounded in 75,309 judgments, paired with LongJud-Bench for cross-stage agent evaluation.
daVinci-kernel is a multi-agent RL system that co-evolves skill selection, policy generation, and summarization via shared LLM and REINFORCE to optimize GPU kernels, reporting higher KernelBench scores than prior RL models.
Proposes COM-as-Action paradigm for deterministic software manipulation, introduces ComCADBench benchmark and ComActor agent that achieves SOTA performance over GUI baselines.
LoopMoE is a looped MoE language model that outperforms matched vanilla MoE on 8 of 9 downstream benchmarks at 3B scale and continues to outperform at 9B scale under strictly controlled budgets.
Stateful visual encoders condition each visual representation on prior features, yielding consistent gains on multi-image tasks under supervised finetuning across model sizes and domains.
OVO-S-Bench provides 1680 human-annotated questions on 348 videos to measure streaming spatial intelligence in MLLMs across instantaneous perception, spatiotemporal tracking, spatial simulation, and allocentric mapping.
Moral Trolley Arena shows frontier LLMs produce composite moral preferences that are compressed rather than additive functions of calibrated component act strengths across Moral Foundations Theory.
VitaBench 2.0 introduces a benchmark for long-term personalized and proactive agent behavior, with results indicating substantial gaps in current frontier LLMs.
LatentOmni proposes a latent-space cross-modal reasoning framework that uses feature-level supervision and Omni-Sync Position Embedding to align and synchronize audio-visual latents, supported by a new 35K interleaved reasoning dataset and showing gains over text CoT baselines.
CopT reverses CoT by eliciting a draft answer first then using continuous-embedding contrastive verification and on-policy thinking to reflect and correct, yielding up to 23% higher accuracy and 57% fewer tokens without training.
PRISM benchmark of over 10k pairs shows LLMs have a 41% average drop from code execution success to spatial correctness in programmatic video generation.
PARAMΔ upcycles dense models to MoE for per-language experts and grafts post-training deltas to enable data-efficient language expansion while preserving original capabilities.
Introduces BacktestBench benchmark with 18k QA pairs across four backtesting tasks and evaluates 23 LLMs via the AutoBacktest multi-agent system.
A genome-conditioned 4B LLM agent predicts microbial life boundaries and matches larger frontier models via token fusion, tool use, and a counterfactual gene-grounding reward.
citing papers explorer
-
LegalWorld: A Life-Cycle Interactive Environment for Legal Agents
LegalWorld is a life-cycle interactive environment modeling Chinese civil litigation as five causally connected stages grounded in 75,309 judgments, paired with LongJud-Bench for cross-stage agent evaluation.
-
Every Act Has Its Price: Compressed Moral Composition in Frontier LLMs
Moral Trolley Arena shows frontier LLMs produce composite moral preferences that are compressed rather than additive functions of calibrated component act strengths across Moral Foundations Theory.
-
LatentOmni: Rethinking Omni-Modal Understanding via Unified Audio-Visual Latent Reasoning
LatentOmni proposes a latent-space cross-modal reasoning framework that uses feature-level supervision and Omni-Sync Position Embedding to align and synchronize audio-visual latents, supported by a new 35K interleaved reasoning dataset and showing gains over text CoT baselines.
-
CopT: Contrastive On-Policy Thinking with Continuous Spaces for General and Agentic Reasoning
CopT reverses CoT by eliciting a draft answer first then using continuous-embedding contrastive verification and on-policy thinking to reflect and correct, yielding up to 23% higher accuracy and 57% fewer tokens without training.
-
A Data-Efficient Path to Multilingual LLMs: Language Expansion via Post-training PARAM$\Delta$ Integration into Upcycled MoE
PARAMΔ upcycles dense models to MoE for per-language experts and grafts post-training deltas to enable data-efficient language expansion while preserving original capabilities.
-
BacktestBench: Benchmarking Large Language Models for Automated Quantitative Strategy Backtesting
Introduces BacktestBench benchmark with 18k QA pairs across four backtesting tasks and evaluates 23 LLMs via the AutoBacktest multi-agent system.
-
StoryAlign: Evaluating and Training Reward Models for Story Generation
StoryReward, trained on a new 100k story preference dataset, sets state-of-the-art performance on the introduced StoryRMB benchmark for aligning LLM stories with human preferences.
-
OmniCompliance-100K: A Multi-Domain, Rule-Grounded, Real-World Safety Compliance Dataset
OmniCompliance-100K supplies 12,985 distinct rules and 106,009 associated real-world cases from 74 multi-domain regulations to benchmark LLM safety and compliance.
-
MemSearcher: Training LLMs to Reason, Search and Manage Memory via End-to-End Reinforcement Learning
MemSearcher trains LLMs to manage compact memory in multi-turn searches via multi-context GRPO for end-to-end RL, outperforming ReAct-style baselines with stable token counts.
-
MSQA: A Natively Sourced Multilingual and Multicultural SimpleQA Benchmark
MSQA benchmark shows LLMs exhibit cultural degradation where competence tracks pre-training data exposure more than reasoning ability, and inference fixes like sampling or retrieval do not close the gap.
-
Toward Generalist Autonomous Research via Hypothesis-Tree Refinement
Arbor combines a coordinator, executors, and a hypothesis tree to enable cumulative autonomous research, outperforming Codex and Claude Code by over 2.5x on six real tasks and reaching 86.36% Any Medal on MLE-Bench Lite.
-
Cross-Source Reasoning-based Correction for Author Name Disambiguation
CrossND is a full-stack framework using data refinement, cross-source reasoning via probabilistic soft logic, and test-time scaling to correct author assignments from inconsistent sources, outperforming 17 baselines on real-world datasets.
-
Self-Evolving Deep Research via Joint Generation and Evaluation
SCORE is a shared-parameter co-evolutionary framework coupling generation and evaluation of deep research reports with a meta-harness to adapt evaluation standards as performance improves.
-
Memory Grafting: Scaling Language Model Pre-training via Offline Conditional Memory
Memory Grafting improves language-model benchmarks by grafting offline hidden-state memory from a larger model into a recipient model using n-gram lookups and lightweight adapters, outperforming MoE and vanilla Engram baselines at 0.92B and 2.8B scales.
-
Stop When Reasoning Converges: Semantic-Preserving Early Exit for Reasoning Models
PUMA detects reasoning-level semantic redundancy to enable early exit in chains of thought, achieving 26.2% average token reduction across five LRMs and five benchmarks while preserving accuracy and CoT quality.
-
Artificial Intolerance: Stigmatizing Language in Clinical Documentation Skews Large Language Model Decision-Making
Frontier LLMs exhibit bias from stigmatizing language in clinical vignettes across four conditions, skewing decisions toward less aggressive management, with limited mitigation from Chain-of-Thought or self-debiasing prompts.
-
Edit-Based Refinement for Parallel Masked Diffusion Language Models
ME-DLM augments parallel masked diffusion models with edit-distance-supervised refinements to raise quality on coding and math benchmarks while using far fewer diffusion steps.
-
Teaching Thinking Models to Reason with Tools: A Full-Pipeline Recipe for Tool-Integrated Reasoning
A training recipe for tool-integrated reasoning models achieves state-of-the-art open-source results on math benchmarks such as 96.7% and 99.2% on AIME 2025 at 4B and 30B scales by balancing tool-use trajectories and optimizing for pass@k during SFT before stable RLVR.
-
MAIC-UI: Making Interactive Courseware with Generative UI
MAIC-UI provides a zero-code authoring system for generating and iteratively editing interactive courseware from educational materials via structured analysis and incremental generation, with lab and classroom evaluations showing usability gains and learning improvements.
-
In-Situ Behavioral Evaluation for LLM Fairness, Not Standardized-Test Scores
Standardized-test benchmarks for LLM fairness are unreliable because prompt wording alone drives most score variance and ranking changes, while a multi-agent conversational framework reveals consistent model-specific fairness behaviors across millions of dialogues.
-
FinReasoning: A Hierarchical Benchmark for Reliable Financial Research Reporting
FinReasoning is a hierarchical benchmark that decomposes LLM financial research capabilities into semantic consistency, data alignment, and deep insight, revealing model-type differences in auditing versus insight generation.
-
BLASST: Dynamic BLocked Attention Sparsity via Softmax Thresholding
BLASST dynamically sparsifies attention by thresholding softmax scores to skip blocks, delivering 1.5x speedups at 70%+ sparsity while preserving benchmark accuracy.
-
MiroThinker: Pushing the Performance Boundaries of Open-Source Research Agents via Model, Context, and Interactive Scaling
MiroThinker shows that scaling agent-environment interactions via reinforcement learning lets a 72B open-source model reach up to 81.9% on GAIA and approach commercial performance on research benchmarks.
-
Value-and-Structure Alignment for Routing-Consistent Quantization of Mixture-of-Experts Models
VSRAQ is a MoE-specific quantization objective that combines value and structure alignment to preserve expert-selection behavior and reduce quality loss without inference overhead.
-
DeepSeek-V3.2: Pushing the Frontier of Open Large Language Models
DeepSeek-V3.2 adds sparse attention, scaled RL post-training, and large-scale agentic data synthesis to reach GPT-5-level performance and gold medals in 2025 IMO and IOI with its high-compute variant.