Presents a new expert-curated dataset of multi-turn counterspeech dialogues in five languages targeting hate against seven groups, with span annotations linking to verified external knowledge for RAG applications.
super hub Mixed citations
Title resolution pending
Mixed citation behavior. Most common role is background (61%).
536
Pith papers citing it
Background
61% of classified citations
abstract
We introduce Mistral 7B v0.1, a 7-billion-parameter language model engineered for superior performance and efficiency. Mistral 7B outperforms Llama 2 13B across all evaluated benchmarks, and Llama 1 34B in reasoning, mathematics, and code generation. Our model leverages grouped-query attention (GQA) for faster inference, coupled with sliding window attention (SWA) to effectively handle sequences of arbitrary length with a reduced inference cost. We also provide a model fine-tuned to follow instructions, Mistral 7B -- Instruct, that surpasses the Llama 2 13B -- Chat model both on human and automated benchmarks. Our models are released under the Apache 2.0 license.
hub tools
citation-role summary
background 57
method 15
baseline 10
other 6
dataset 2
citation-polarity summary
claims ledger
- abstract We introduce Mistral 7B v0.1, a 7-billion-parameter language model engineered for superior performance and efficiency. Mistral 7B outperforms Llama 2 13B across all evaluated benchmarks, and Llama 1 34B in reasoning, mathematics, and code generation. Our model leverages grouped-query attention (GQA) for faster inference, coupled with sliding window attention (SWA) to effectively handle sequences of arbitrary length with a reduced inference cost. We also provide a model fine-tuned to follow instructions, Mistral 7B -- Instruct, that surpasses the Llama 2 13B -- Chat model both on human and auto
authors
co-cited works
representative citing papers
Introduces BonaFide benchmark of 3,066 ground-truth labeled CoTs showing most faithfulness metrics perform near chance with biases and poor scaling to longer chains.
RTI-Bench is the first publicly released structured dataset of CIC administrative decisions with outcome labels, exemption citations, IRAC reasoning, and timelines, built from 1,218 corpus cases and 298 PDFs, achieving 95.3% label precision on manual review and 57.3% accuracy on a Mistral 7B zero-Sh
Zero-Run auditing supplies valid lower bounds on differential privacy parameters from fixed member and non-member datasets by modeling and correcting distribution-shift confounding via causal-inference techniques.
LCDD creates sparse carriers for SFT behaviors that SFT-Eraser can reverse, with ablations showing the sparse structure enables causal control.
INT4 quantization recovers up to 22 times more forgotten training data in unlearned LLMs, and the proposed DURABLEUN-SAF method is the first to maintain forgetting across BF16, INT8, and INT4 precisions.
An adversary controlling an intermediate pipeline stage in decentralized LLM post-training can inject a backdoor that reduces alignment from 80% to 6%, with the backdoor persisting in 60% of cases even after subsequent safety training.
CacheTrap achieves 100% targeted attack success on five open-source LLMs by using an efficient search to locate and flip a single bit in the KV cache as a transient trigger, while preserving normal accuracy without the trigger.
MediQAl is a new French medical QA benchmark with 32k exam-sourced questions in three formats and cognitive labels, evaluated on 14 LLMs to reveal gaps between factual recall and reasoning performance.
LLaDA is a scalable diffusion-based language model that matches autoregressive LLMs like LLaMA3 8B on tasks and surpasses GPT-4o on reversal poem completion.
MME-RealWorld is the largest manually annotated high-resolution benchmark for MLLMs, where even the best models achieve less than 60% accuracy on challenging real-world tasks.
LiveBench is a contamination-limited LLM benchmark with auto-scored challenging tasks from recent sources across math, coding, reasoning and more, where top models score below 70%.
ORPO performs preference alignment during supervised fine-tuning via a monolithic odds ratio penalty, allowing 7B models to outperform larger state-of-the-art models on alignment benchmarks.
Creates LoCoMo benchmark dataset for very long-term LLM conversational memory and shows current models struggle with lengthy dialogues and long-range temporal dynamics.
The paper defines STE and SPID, two information-theoretic measures of semantic flow and decomposition in language exchanges, and applies them to four dialogue datasets.
Anisotropy, quantified by dominant-dimension variance fraction, determines the best parameter-free similarity metric for text embeddings, with rank-based metrics gaining ~20% relative where cosine is weakest.
MultiHashFormer enables hash-based autoregression in LMs by encoding tokens as multi-hash signatures, outperforming standard Transformers at 100M-3B scales while keeping parameter count constant for multilingual expansion.
Introduces nexbax, a diagnostic framework with three themes and 10 dimensions for evaluating AI economic viability, operational practicality, and societal integrity in next-billion-user contexts.
A reference-based geometric hashing method recovers cross-model vector correspondences by exploiting local isometric consistency in contrastive embeddings and iteratively bootstrapping from a seed of paired anchors.
The study links three LVLM architectural dimensions to three hallucination types via a new benchmark, finding that language foundation quality reduces co-occurrence errors, visual encoder strength reduces similarity errors, alignment reduces uncertainty errors, and joint visual-alignment improvement
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.
Introduces SANSA paradigm for semantic-agnostic vision-language segmentation via dictionary or example-based prompts, with finetuning delivering up to 20% mIoU gains on the new task while retaining standard performance.
MentalMap benchmark identifies a universal L3 reasoning cliff in LLMs' text-based spatial reasoning that persists across languages, scales, and prompting, and is replicated in human evaluations.
MATCHA introduces a dual-view contrastive metric measuring proximity to gold text and distance from adversarial contradictions, outperforming ROUGE and BERTScore by up to 20% on TruthfulQA and other NLP benchmarks.
citing papers explorer
-
SpikingBrain2.0: Brain-Inspired Foundation Models for Efficient Long-Context and Cross-Platform Inference
SpikingBrain2.0 is a 5B hybrid spiking-Transformer that recovers most base model performance while delivering 10x TTFT speedup at 4M context and supporting over 10M tokens on limited GPUs via dual sparse attention and dual quantization paths.
-
A Sociotechnical, Practitioner-Centered Approach to Technology Adoption in Cybersecurity Operations: An LLM Case
A six-month ethnographic co-creation project in a real SOC demonstrates that practitioner involvement in LLM tool design can overcome typical adoption barriers in cybersecurity operations.
-
GS-Quant: Granular Semantic and Generative Structural Quantization for Knowledge Graph Completion
GS-Quant generates coarse-to-fine discrete codes for KG entities via semantic hierarchy injection and causal sequence reconstruction, enabling LLMs to perform knowledge graph completion by treating the codes as vocabulary tokens.
-
ReCAPA: Hierarchical Predictive Correction to Mitigate Cascading Failures
ReCAPA adds predictive correction and multi-level semantic alignment to VLA models, plus two new metrics for tracking error spread and recovery, yielding competitive benchmark results over LLM baselines.
-
Nexusformer: Nonlinear Attention Expansion for Stable and Inheritable Transformer Scaling
Nexusformer uses a three-stage nonlinear mapping in attention to enable stable, inheritable scaling of transformers, matching baseline perplexity with up to 41.5% less compute when growing from 240M to 440M parameters.
-
Towards Scalable Lifelong Knowledge Editing with Selective Knowledge Suppression
LightEdit enables scalable lifelong knowledge editing in LLMs via selective knowledge retrieval and probability suppression during decoding, outperforming prior methods on ZSRE, Counterfact, and RIPE while reducing training costs.
-
STK-Adapter: Incorporating Evolving Graph and Event Chain for Temporal Knowledge Graph Extrapolation
STK-Adapter adds Spatial-Temporal MoE, Event-Aware MoE, and Cross-Modality Alignment MoE to integrate evolving TKG graphs and event chains into LLMs, reducing information loss and improving extrapolation performance over prior methods.
-
TabEmb: Joint Semantic-Structure Embedding for Table Annotation
TabEmb decouples LLM-based semantic column embeddings from graph-based structural modeling to produce joint representations that improve table annotation tasks.
-
Handling and Interpreting Missing Modalities in Patient Clinical Trajectories via Autoregressive Sequence Modeling
Autoregressive transformer modeling with missingness-aware contrastive pre-training outperforms baselines on MIMIC-IV and eICU benchmarks and mitigates divergent behavior from removed modalities in clinical trajectories.
-
Medical Image Understanding Improves Survival Prediction via Visual Instruction Tuning
A vision-language model pre-trained via instruction tuning on CT-report pairs improves survival prediction accuracy over baselines, especially when clinical data alone is weak, while also producing text answers to clinical questions.
-
Linear-Time and Constant-Memory Text Embeddings Based on Recurrent Language Models
Fine-tuned recurrent models like Mamba2 produce competitive text embeddings with linear-time constant-memory inference via vertical chunking, outperforming transformers in memory use.
-
Align Documents to Questions: Question-Oriented Document Rewriting for Retrieval-Augmented Generation
QREAM rewrites documents to question-focused style using iterative ICL and distilled FT models, boosting RAG performance by up to 8% relative improvement.
-
Calibrating Model-Based Evaluation Metrics for Summarization
A reference-free proxy scoring framework combined with GIRB calibration produces better-aligned evaluation metrics for summarization and outperforms baselines across seven datasets.
-
TStore: Rethinking AI Model Hub with Tensor-Centric Compression
TStore reduces AI model storage via tensor-level fingerprinting, clustering, and compression without annotations while claiming to preserve usability.
-
HieraSparse: Hierarchical Semi-Structured Sparse KV Attention
HieraSparse delivers a hierarchical semi-structured sparse KV attention system that achieves 1.2x KV compression and 4.57x decode attention speedup versus prior unstructured sparsity methods at equivalent sparsity, plus up to 1.85x prefill speedup and 1.37x/1.77x speedups with magnitude pruning and
-
Pruning Unsafe Tickets: A Resource-Efficient Framework for Safer and More Robust LLMs
Pruning removes 'unsafe tickets' from LLMs via gradient-free attribution, reducing harmful outputs and jailbreak vulnerability with minimal utility loss.
-
Rethinking the Necessity of Adaptive Retrieval-Augmented Generation through the Lens of Adaptive Listwise Ranking
AdaRankLLM shows adaptive listwise reranking outperforms fixed-depth retrieval for most LLMs by acting as a noise filter for weak models and an efficiency optimizer for strong ones, with lower context use.
-
IUQ: Interrogative Uncertainty Quantification for Long-Form Large Language Model Generation
IUQ quantifies claim-level uncertainty in long-form LLM generation by combining inter-sample consistency and intra-sample faithfulness through an interrogate-then-respond approach and outperforms baselines on two datasets.
-
Psychological Concept Neurons: Can Neural Control Bias Probing and Shift Generation in LLMs?
Big Five personality traits become decodable early in LLMs, are represented by mid-layer selective neurons, and can be shifted by targeted activation interventions, though effects on generated labels are weaker and spill across traits.
-
Please Make it Sound like Human: Encoder-Decoder vs. Decoder-Only Transformers for AI-to-Human Text Style Transfer
BART-large outperforms Mistral-7B in AI-to-human style transfer with higher reference similarity scores and far fewer parameters, while showing that marker shift can reflect overshoot rather than accurate transfer.
-
FAITH: Factuality Alignment through Integrating Trustworthiness and Honestness
FAITH improves LLM factual accuracy by mapping confidence and semantic entropy into natural-language knowledge-state quadrants for trustworthiness and honestness, then applying PPO with a combined reward and retrieval augmentation.
-
Adapt to Thrive! Adaptive Power-Mean Policy Optimization for Improved LLM Reasoning
APMPO boosts average Pass@1 scores on math reasoning benchmarks by 3 points over GRPO by using an adaptive power-mean policy objective and feedback-driven clipping bounds in RLVR training.
-
Free Energy-Driven Reinforcement Learning with Adaptive Advantage Shaping for Unsupervised Reasoning in LLMs
FREIA applies free energy principles and adaptive advantage shaping to unsupervised RL, outperforming baselines by 0.5-3.5 Pass@1 points on math reasoning with a 1.5B model.
-
Policy-Aware Edge LLM-RAG Framework for Internet of Battlefield Things Mission Orchestration
PA-LLM-RAG adds policy retrieval and dual-LLM verification to enable reliable low-latency mission orchestration in simulated IoBT environments, with Gemma-2B reaching 100% policy compliance at 4.17s latency.
-
LanG -- A Governance-Aware Agentic AI Platform for Unified Security Operations
LanG presents a governance-aware agentic AI platform for unified security operations that reports strong performance on incident correlation, rule generation, attack reconstruction, and AI safety guardrails in an open-source package.
-
PassiveQA: A Three-Action Framework for Epistemically Calibrated Question Answering via Supervised Finetuning
PassiveQA trains models via supervised finetuning to decide Answer, Ask, or Abstain using structured information-state representations and knowledge-graph context, yielding better abstention and lower hallucination on QA datasets.
-
SUMMIR: A Hallucination-Aware Framework for Ranking Sports Insights from LLMs
SUMMIR is a multimetric ranking model that orders LLM-generated sports insights by importance while incorporating hallucination detection to improve factual reliability across cricket, soccer, basketball, and baseball articles.
-
Tug-of-War within A Decade: Conflict Resolution in Vulnerability Analysis via Teacher-Guided Retrieval-Augmented Generations
CRVA-TGRAG combines parent-document segmentation, ensemble retrieval, and teacher-guided fine-tuning to mitigate knowledge conflicts and improve accuracy in LLM-based CVE vulnerability analysis.
-
Benchmarking Linguistic Adaptation in Comparable-Sized LLMs: A Study of Llama-3.1-8B, Mistral-7B-v0.1, and Qwen3-8B on Romanized Nepali
Fine-tuning Llama-3.1-8B, Mistral-7B-v0.1, and Qwen3-8B on Romanized Nepali data enables effective generation where zero-shot fails, with Qwen3-8B performing best overall and Llama-3.1-8B showing the largest gains.
-
Explicit Logic Channel for Validation and Enhancement of MLLMs on Zero-Shot Tasks
Introduces Explicit Logic Channel (ELC) with LLM, VFM and probabilistic inference for validating, selecting and enhancing MLLMs on zero-shot tasks using Consistency Rate and cross-channel integration.
-
On the Limits of Layer Pruning for Generative Reasoning in Large Language Models
Layer pruning preserves classification performance in LLMs but fundamentally limits recovery of generative reasoning capabilities even after extensive self-supervised finetuning.
-
Different types of syntactic agreement recruit the same units within large language models
Different types of syntactic agreement recruit overlapping units within LLMs, indicating that agreement forms a meaningful functional category across English, Russian, Chinese, and structurally similar languages.
-
Chinese Short-Form Creative Content Generation via Explanation-Oriented Multi-Objective Optimization
MAGIC-HMO is a multi-agent framework that treats Chinese short-form creative NLG as heterogeneous multi-objective optimization over personalized constraints plus explanation reliability and outperforms baselines on a baby-naming benchmark.
-
LLM4Delay: Flight Delay Prediction via Cross-Modality Adaptation of Large Language Models and Aircraft Trajectory Representation
LLM4Delay improves flight delay prediction accuracy by using instance-level projection to adapt LLMs for integrating textual aeronautical information with multiple aircraft trajectories.
-
Retrofitting Small Multilingual Models for Retrieval: Matching 7B Performance with 300M Parameters
A 300M multilingual embedding model matches or exceeds 7B retrieval performance via optimized data scale, hard negatives, and task diversity over language diversity.
-
SpikingMamba: Towards Energy-Efficient Large Language Models via Knowledge Distillation from Mamba
SpikingMamba distills Mamba into an SNN LLM achieving 4.76x energy savings with a 4.78% zero-shot accuracy gap that narrows to 2.23% after RL.
-
OntoLogX: Ontology-Guided Knowledge Graph Extraction from Cybersecurity Logs with Large Language Models
OntoLogX is a system that applies LLMs with ontology guidance, RAG, and iterative fixes to build valid knowledge graphs from cybersecurity logs and predict ATT&CK tactics from aggregated sessions.
-
BoHA: Blockwise Hadamard Product Adaptation for Parameter-Efficient Fine-Tuning
BoHA partitions frozen weights into a b by b grid and applies independent low-rank Hadamard factors per block, outperforming LoRA on matched-budget single-task averages while retaining 57.66% first-stage accuracy in a commonsense-to-arithmetic continual-learning test on Llama-3.2-3B.
-
ShadowNPU: System and Algorithm Co-design for NPU-Centric On-Device LLM Inference
ShadowNPU presents shadowAttn, a co-designed sparse attention system that uses NPU pilot compute and techniques like graph bucketing and per-head sparsity to minimize CPU/GPU fallback during on-device LLM inference while maintaining accuracy.
-
Mitigating Watermark Forgery in Generative Models via Randomized Key Selection
Randomized per-query key selection with single-key detection acceptance bounds forgery success rate independently of collected samples while preserving model utility.
-
From 2:4 to 8:16 sparsity patterns in LLMs for Outliers and Weights with Variance Correction
8:16 sparsity with variance correction and outlier handling lets compressed LLMs match or exceed dense-model accuracy under fixed memory limits, outperforming the common 2:4 pattern in flexibility.
-
A Survey on Vision-Language-Action Models: An Action Tokenization Perspective
The survey frames VLA models as pipelines that generate progressively grounded action tokens and classifies those tokens into eight types to guide future development.
-
Task-conditioned probing of instruction-tuned multimodal LLMs: Region-specific brain alignment patterns under naturalistic stimuli
Instruction-tuned multimodal LLMs show higher brain alignment (~9-20% over baselines) under task-specific instructions during naturalistic video stimuli, with task-conditioned subspaces and weak semantic coupling.
-
Enhancing the Safety of Medical Vision-Language Models by Synthetic Demonstrations
Synthetic clinical demonstrations at inference time improve safety of Med-VLMs against visual and textual jailbreaks while preserving general performance on medical tasks.
-
Through the Stealth Lens: Attention-Aware Defenses Against Poisoning in RAG
Introduces NPAS and AV Filter using LLM attention weights to defend RAG against poisoning, reporting up to 20% accuracy gains while adaptive attacks reach 35% success.
-
Beyond Words: Multimodal LLM Knows When to Speak
MM-When2Speak reformulates conversational timing as dense response-type prediction and achieves up to 3x better performance by integrating video, audio, and text cues on top of an LLM backbone using a new dyadic conversation dataset.
-
Advancing AI Research Assistants with Expert-Involved Learning
ARIEL evaluates LLMs and LMMs on full-length biomedical summarization and figure interpretation with blinded expert review, identifies limitations, and demonstrates gains from prompt engineering, fine-tuning, and an integrated agent for hypothesis generation.
-
Low-Confidence Gold: Refining Low-Confidence Samples for Efficient Instruction Tuning
LCG applies centroid clustering and confidence-guided semi-supervised selection to curate 6K-sample subsets that yield superior MT-bench performance compared with prior filtering methods when used for instruction tuning.
-
MSMO-ABSA: Multi-Scale and Multi-Objective Optimization for Cross-Lingual Aspect-Based Sentiment Analysis
MSMO framework achieves claimed SOTA cross-lingual ABSA via sentence- and aspect-level alignment, code-switching, consistency training, and knowledge distillation.
-
Social media polarization during conflict: Insights from an ideological stance dataset on Israel-Palestine Reddit comments
A new labeled dataset of 9,969 Israel-Palestine Reddit comments is created and used to compare stance classification methods, with a specific Mixtral prompt achieving the highest performance.