BEHAVIOR-1K introduces a benchmark of 1,000 human everyday activities in realistic simulated scenes together with the OMNIGIBSON physics simulator to evaluate embodied AI.
hub Mixed citations
GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding
Mixed citation behavior. Most common role is background (57%).
74
Pith papers citing it
Background
57% of classified citations
abstract
For natural language understanding (NLU) technology to be maximally useful, both practically and as a scientific object of study, it must be general: it must be able to process language in a way that is not exclusively tailored to any one specific task or dataset. In pursuit of this objective, we introduce the General Language Understanding Evaluation benchmark (GLUE), a tool for evaluating and analyzing the performance of models across a diverse range of existing NLU tasks. GLUE is model-agnostic, but it incentivizes sharing knowledge across tasks because certain tasks have very limited training data. We further provide a hand-crafted diagnostic test suite that enables detailed linguistic analysis of NLU models. We evaluate baselines based on current methods for multi-task and transfer learning and find that they do not immediately give substantial improvements over the aggregate performance of training a separate model per task, indicating room for improvement in developing general and robust NLU systems.
hub tools
citation-role summary
background 7
dataset 7
citation-polarity summary
claims ledger
- abstract For natural language understanding (NLU) technology to be maximally useful, both practically and as a scientific object of study, it must be general: it must be able to process language in a way that is not exclusively tailored to any one specific task or dataset. In pursuit of this objective, we introduce the General Language Understanding Evaluation benchmark (GLUE), a tool for evaluating and analyzing the performance of models across a diverse range of existing NLU tasks. GLUE is model-agnostic, but it incentivizes sharing knowledge across tasks because certain tasks have very limited train
co-cited works
representative citing papers
LIBERO is a new benchmark for lifelong robot learning that evaluates transfer of declarative, procedural, and mixed knowledge across 130 manipulation tasks with provided demonstration data.
Task vectors from weight differences allow arithmetic operations to edit pre-trained models, improving multiple tasks simultaneously and enabling analogical inference on unseen tasks.
An unsupervised technique extracts latent yes-no knowledge from language model activations by locating a direction that satisfies logical consistency properties, outperforming zero-shot accuracy by 4% on average across models and datasets.
Introduces the MMLU benchmark of 57 tasks and shows that current models, including GPT-3, achieve low accuracy far below expert level across academic and professional domains.
A new probing framework detects moderate parametric memorization signals in tabular in-context learning models under single-task fine-tuning, strongest on low-cardinality tasks, but signals largely disappear under realistic training.
DICE formalizes multi-agent LLM coordination as discounted incomplete-information Markov games and introduces Heterogeneous Quantal Response Equilibrium (HQRE) to achieve unique stable equilibria with bounded regret, demonstrated via prompt-control and fine-tuning algorithms on eleven benchmarks.
EpiCastBench supplies 40 curated multivariate epidemic datasets and evaluates 15 forecasting models under unified preprocessing, horizons, metrics, and significance tests.
Intrinsic Muon provides closed-form linear maximization oracles on multiple Riemannian matrix manifolds for unitarily invariant norms, with convergence rates depending only on manifold dimension or rank.
Evolutionary trees from LLM weights recover ground-truth training topologies and identify key datasets and layers through phenotypic analysis.
MIXAR is the first autoregressive pixel-based language model for eight languages and scripts, with empirical gains on multilingual tasks, robustness to unseen languages, and further improvements when scaled to 0.5B parameters.
Winner-take-all spiking self-attention replaces softmax in spiking transformers to support language modeling on 16 datasets with spike-driven, energy-efficient architectures.
LoRA weight updates are spectrally sparse, with 33% of DCT coefficients capturing 90% of energy on average, enabling 10x storage reduction and occasional gains by masking high frequencies.
Norm-Anchor Scaling breaks the norm-feedback loop in sequential LLM editing by anchoring value vectors to original norms, improving long-run performance by 72.2% and extending the editing horizon over 4x.
UWE is a task-agnostic bi-encoder that uses many-to-many InfoNCE and token-level soft late interaction to achieve zero-shot ranking across unseen work-related target spaces while using far fewer parameters than Qwen3-8B and improving MAP by 4.4 points.
Power-Softmax is a new HE-compatible attention variant that permits training and inference of billion-parameter polynomial LLMs with performance matching standard transformers.
K-sparse autoencoders with dead-latent fixes produce clean scaling laws and better feature quality metrics that improve with size, shown by training a 16-million-latent model on GPT-4 activations.
QLoRA finetunes 4-bit quantized LLMs via LoRA adapters to match full-precision performance while using far less memory, enabling 65B-scale training on single GPUs and producing Guanaco models near ChatGPT level.
LLaMA-Adapter turns frozen LLaMA 7B into a capable instruction follower using only 1.2M new parameters and zero-init attention, matching Alpaca while extending to image-conditioned reasoning on ScienceQA and COCO.
LLM.int8() performs 8-bit inference for transformers up to 175B parameters with no accuracy loss by combining vector-wise quantization for most features with 16-bit mixed-precision handling of systematic outlier dimensions.
Switch Transformers use top-1 expert routing in a Mixture of Experts setup to scale to trillion-parameter language models with constant compute and up to 4x speedup over T5-XXL.
XLM-R, pretrained on 100 languages with 2TB of CommonCrawl data, improves average XNLI accuracy by 14.6 points and MLQA F1 by 13 points over mBERT while matching strong monolingual models on GLUE.
BART introduces a denoising pretraining method for seq2seq models that matches RoBERTa on GLUE and SQuAD while setting new state-of-the-art results on abstractive summarization, dialogue, and QA with up to 6 ROUGE gains.
T5 casts all NLP tasks as text-to-text generation, systematically explores pre-training choices, and reaches strong performance on summarization, QA, classification and other tasks via large-scale training on the Colossal Clean Crawled Corpus.
citing papers explorer
-
Probing Memorization of Tabular In-Context Learning
A new probing framework detects moderate parametric memorization signals in tabular in-context learning models under single-task fine-tuning, strongest on low-cardinality tasks, but signals largely disappear under realistic training.
-
DICE: Entropy-Regularized Equilibrium Selection for Stable Multi-Agent LLM Coordination
DICE formalizes multi-agent LLM coordination as discounted incomplete-information Markov games and introduces Heterogeneous Quantal Response Equilibrium (HQRE) to achieve unique stable equilibria with bounded regret, demonstrated via prompt-control and fine-tuning algorithms on eleven benchmarks.
-
EpiCastBench: Datasets and Benchmarks for Multivariate Epidemic Forecasting
EpiCastBench supplies 40 curated multivariate epidemic datasets and evaluates 15 forecasting models under unified preprocessing, horizons, metrics, and significance tests.
-
Intrinsic Muon: Spectral Optimization on Riemannian Matrix Manifolds
Intrinsic Muon provides closed-form linear maximization oracles on multiple Riemannian matrix manifolds for unitarily invariant norms, with convergence rates depending only on manifold dimension or rank.
-
Analysis and Explainability of LLMs Via Evolutionary Methods
Evolutionary trees from LLM weights recover ground-truth training topologies and identify key datasets and layers through phenotypic analysis.
-
MIXAR: Scaling Autoregressive Pixel-based Language Models to Multiple Languages and Scripts
MIXAR is the first autoregressive pixel-based language model for eight languages and scripts, with empirical gains on multilingual tasks, robustness to unseen languages, and further improvements when scaled to 0.5B parameters.
-
Winner-Take-All Spiking Transformer for Language Modeling
Winner-take-all spiking self-attention replaces softmax in spiking transformers to support language modeling on 16 datasets with spike-driven, energy-efficient architectures.
-
SpectralLoRA: Is Low-Frequency Structure Sufficient for LoRA Adaptation? A Spectral Analysis of Weight Updates
LoRA weight updates are spectrally sparse, with 33% of DCT coefficients capturing 90% of energy on average, enabling 10x storage reduction and occasional gains by masking high frequencies.
-
Norm Anchors Make Model Edits Last
Norm-Anchor Scaling breaks the norm-feedback loop in sequential LLM editing by anchoring value vectors to original norms, improving long-run performance by 72.2% and extending the editing horizon over 4x.
-
STAR: Rethinking MoE Routing as Structure-Aware Subspace Learning
STAR rethinks MoE routing as structure-aware subspace learning by adding a GHA-tracked principal subspace to standard routers, yielding more stable specialization and better performance on synthetic, language, and vision tasks.
-
Towards Evaluation Engineering: An Empirical Study of ML Evaluation Harnesses in the Wild
An empirical study of 57 ML evaluation harnesses shows 41.4% of operational issues occur in the specification stage, driven mainly by unimplemented features, documentation gaps, and missing input validation.
-
UniSpike: Accelerating Spiking Neural Networks on Neuromorphic Systems via Eliminating Address Redundancy
UniSpike eliminates address redundancy in spike packets via co-design of scheduling, runtime assembly hardware, and SNN partitioning, reporting 1.93x average traffic reduction, 1.77x speedup, and 1.50x energy improvement.
-
ZeroUnlearn: Few-Shot Knowledge Unlearning in Large Language Models
ZeroUnlearn reformulates machine unlearning as knowledge re-mapping via model editing, using multiplicative updates with closed-form solutions for efficient few-shot removal of sensitive representations while preserving utility.
-
PEML: Parameter-efficient Multi-Task Learning with Optimized Continuous Prompts
PEML co-optimizes continuous prompts and low-rank adaptations to deliver up to 6.67% average accuracy gains over existing multi-task PEFT methods on GLUE, SuperGLUE, and other benchmarks.
-
SURGE: Surrogate Gradient Adaptation in Binary Neural Networks
SURGE proposes a dual-path gradient compensator and adaptive gradient scaler to mitigate gradient mismatch in binary neural network training via auxiliary backpropagation.
-
AdaPreLoRA: Adafactor Preconditioned Low-Rank Adaptation
AdaPreLoRA pairs the Adafactor diagonal Kronecker preconditioner on the full weight matrix with a closed-form factor-space solve that selects the update minimizing an H_t-weighted imbalance, yielding competitive results on GPT-2, Mistral-7B, Qwen2-7B and diffusion personalization tasks.
-
PragLocker: Protecting Agent Intellectual Property in Untrusted Deployments via Non-Portable Prompts
PragLocker generates function-preserving but non-portable prompts for LLM agents via code-symbol semantic anchoring followed by target-model feedback noise injection.
-
Finding Meaning in Embeddings: Concept Separation Curves
Concept Separation Curves provide a classifier-independent method to visualize and quantify how sentence embeddings distinguish conceptual meaning from syntactic variations across languages and domains.
-
TLoRA: Task-aware Low Rank Adaptation of Large Language Models
TLoRA jointly optimizes LoRA initialization via task-data SVD and sensitivity-driven rank allocation, delivering stronger results than standard LoRA across NLU, reasoning, math, code, and chat tasks while using fewer trainable parameters.
-
Memory-Efficient Transfer Learning with Fading Side Networks via Masked Dual Path Distillation
MDPD mutually distills knowledge between a frozen backbone and a learnable side network during fine-tuning, then discards the side network at inference to accelerate speed by at least 25% while preserving accuracy.
-
MP-ISMoE: Mixed-Precision Interactive Side Mixture-of-Experts for Efficient Transfer Learning
MP-ISMoE uses Gaussian noise perturbed iterative quantization and interactive side mixture-of-experts to deliver higher accuracy than prior memory-efficient transfer learning methods while keeping similar parameter and memory usage.
-
The FIL Hypothesis: Inductive Biases Help with Kernel Engineering
The FIL Hypothesis claims that inductive biases outperform purely data-driven methods on GPU programming tasks with non-trivial feedback loops.
-
Steering LLM Viewpoints through Fabricated Evidence Injection
Ghostwriter attack injects fabricated evidence to steer LLM viewpoints, with experiments showing high success on commercial models and partial mitigation on guarded ones.
-
ReLoRA: Knowledge-Reusing Adaptation for Fast Rollout of Evolving LLM Services
ReLoRA reduces time-to-readiness for LoRA adapters on updated LLMs by up to 8.9x through adaptive Bayesian initialization and scheduled regularization while improving accuracy by up to 4.6%.
-
Kernel-Based ReLU Approximation for Homomorphic Encryption-Compatible Privacy-preserving Deep Learning Models
Kernel-based ReLU is approximated by a quadratic polynomial for low-depth homomorphic encryption compatibility, trained on LLM token embeddings and evaluated across DL and transformer settings.
-
Interactive Evaluation Requires a Design Science
Interactive evaluation of AI must be reframed as a distinct paradigm that maps interaction trajectories to judgments on process, recoverability, coordination, robustness, and system performance, supported by a two-axis taxonomy and design principles.
-
Convex Dataset Valuation for Post-Training
A convex KMM-based valuation method that accounts for both target-task alignment and inter-dataset redundancy in gradient space outperforms standard gradient-alignment baselines for LLM post-training data selection.
-
Strategic Over-Parameterization for Generalizable Low-Rank Adaptation
LoRA-Over injects auxiliary parameters into low-rank adapters during training and decomposes them back into standard LoRA at inference, with static or dynamic scheduling to allocate extra capacity where needed, yielding better generalization than vanilla LoRA on GLUE, MT-Bench, GSM8K and HumanEval.
-
Hardware-Efficient Softmax and Layer Normalization with Guaranteed Normalization for Edge Devices
Hardware approximations for Softmax and LayerNorm preserve exact normalization guarantees and deliver up to 14x area reduction in 28nm silicon with negligible accuracy loss on GLUE, SQuAD, and perplexity.
-
A KL Lens on Quantization: Fast, Forward-Only Sensitivity for Mixed-Precision SSM-Transformer Models
KL divergence provides a superior forward-only metric for identifying quantization-sensitive parts in SSM-Transformer hybrids, outperforming MSE and SQNR and supporting practical mixed-precision deployment on edge devices.
-
Adaptive Spiking Neurons for Vision and Language Modeling
ASN uses trainable parameters for adaptive membrane dynamics and firing in SNNs, with NASN adding normalization, and reports effectiveness across 19 vision and language datasets.
-
BWTA: Accurate and Efficient Binarized Transformer by Algorithm-Hardware Co-design
BWTA achieves near full-precision accuracy on BERT and LLMs using binary weights and ternary activations, with 16-24x kernel speedups via specialized CUDA kernels.
-
Sparse-on-Dense: Area and Energy-Efficient Computing of Sparse Neural Networks on Dense Matrix Multiplication Accelerators
Sparse neural networks achieve better area and energy efficiency when executed on dense matrix multiplication accelerators using a Sparse-on-Dense approach than on dedicated sparse accelerators.
-
Transformer Scalability Crisis: The First Comprehensive Empirical Analysis of Performance Walls in Modern Language Models
Empirical tests on 118 transformers show success falling from 88.1% at 512 tokens to 0% at 2048 tokens, with compressed models achieving 649.2 tokens/sec/M parameters versus 12.5 for large generative ones.