BLaIR is a new benchmark and 570M-review dataset showing that LLM performance rankings on recommendation tasks have little correlation with rankings on general embedding benchmarks like MTEB.
super hub Mixed citations
RoBERTa: A Robustly Optimized BERT Pretraining Approach
Mixed citation behavior. Most common role is background (63%).
395
Pith papers citing it
Background
63% of classified citations
abstract
Language model pretraining has led to significant performance gains but careful comparison between different approaches is challenging. Training is computationally expensive, often done on private datasets of different sizes, and, as we will show, hyperparameter choices have significant impact on the final results. We present a replication study of BERT pretraining (Devlin et al., 2019) that carefully measures the impact of many key hyperparameters and training data size. We find that BERT was significantly undertrained, and can match or exceed the performance of every model published after it. Our best model achieves state-of-the-art results on GLUE, RACE and SQuAD. These results highlight the importance of previously overlooked design choices, and raise questions about the source of recently reported improvements. We release our models and code.
hub tools
citation-role summary
background 45
method 12
baseline 5
dataset 3
citation-polarity summary
claims ledger
- abstract Language model pretraining has led to significant performance gains but careful comparison between different approaches is challenging. Training is computationally expensive, often done on private datasets of different sizes, and, as we will show, hyperparameter choices have significant impact on the final results. We present a replication study of BERT pretraining (Devlin et al., 2019) that carefully measures the impact of many key hyperparameters and training data size. We find that BERT was significantly undertrained, and can match or exceed the performance of every model published after it
authors
co-cited works
representative citing papers
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.
Randomly replacing labels in in-context demonstrations barely hurts performance, showing that label space, input distribution, and sequence format drive in-context learning more than ground-truth labels.
SimCSE achieves 76.3% unsupervised and 81.6% supervised Spearman's correlation on STS tasks with BERT-base, improving prior best results by 4.2% and 2.2% via simple contrastive learning.
The Pile is a newly constructed 825 GiB dataset from 22 diverse sources that enables language models to achieve better performance on academic, professional, and cross-domain tasks than models trained on Common Crawl variants.
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.
GPT-3 shows that scaling an autoregressive language model to 175 billion parameters enables strong few-shot performance across diverse NLP tasks via in-context prompting without fine-tuning.
ELECTRA replaces masked language modeling with replaced token detection, yielding contextual representations that outperform BERT at equal compute and match larger models like RoBERTa with far less compute.
REALM augments language-model pre-training with an unsupervised retriever over Wikipedia documents and reports 4-16% absolute gains on open-domain QA benchmarks over prior implicit and explicit knowledge methods.
Sentence-BERT adapts BERT with siamese and triplet networks to produce sentence embeddings for efficient cosine-similarity comparisons, cutting computation time from hours to seconds on similarity search while matching BERT accuracy.
An RL-guided MCTS proof search for Tamarin finds more and shorter proofs than standard search across 16 protocol models.
RISE is an inference-time semantic reranking framework that refines low-confidence predictions in rhetorical role labeling using contrastively learned label representations, delivering an average +9.15 macro-F1 gain on hard examples across eight datasets and seven models.
A solvable hierarchical model with power-law feature strengths yields explicit power-law scaling of prediction error through sequential recovery of latent directions by a layer-wise spectral algorithm.
BOOKMARKS introduces searchable bookmarks as reusable answers to storyline questions, enabling active initialization and passive synchronization for more consistent role-playing agent memory than recurrent summarization.
LLMs copy biased analyst ratings in investment decisions but a new detection method encourages independent reasoning and can improve stock return predictions beyond human levels.
Chain-based Distillation constructs a sequence of anchor models to enable efficient initialization of variable-sized SLMs through interpolation, with bridge distillation for cross-architecture transfer, yielding better performance than scratch training.
A Dutch BERT model encodes gender linearly by epoch 20 but does not dynamically update its representations when explicit female cues contradict learned stereotypical associations in short sentence templates.
GLoRA replaces raw factor averaging with gauge-aware aggregation in a consensus subspace estimated from client projectors, enabling consistent low-rank federated LoRA under heterogeneity.
Code LLMs generate substantially worse comments outside English, and no tested automatic metric or LLM judge reliably matches human assessment of those outputs.
AS-LoRA adaptively chooses which LoRA factor to update per layer and round using a curvature-aware second-order score, eliminating reconstruction error floors and improving performance in DP federated learning.
LLMs outperform single human raters at spotting relative weaknesses in L2 writing profiles on the ICNALE GRA dataset while humans are better at spotting strengths, using a self-referential intra-learner evaluation method.
Pretrained language models are used as energy functions for Glauber dynamics in discrete text diffusion, improving generation quality over prior diffusion LMs and matching autoregressive models on benchmarks and reasoning tasks.
CGFuse enables deep token-level fusion of graph-derived structural features into language models, yielding 10-16% BLEU and 6-11% CodeBLEU gains on code generation tasks.
MedStruct-S benchmark shows encoder-only models outperform larger decoder-only ones on key-conditioned QA from noisy OCR clinical reports, with fine-tuned large models winning only when scale is ignored.
citing papers explorer
-
Bridging Language and Items for Retrieval and Recommendation: Benchmarking LLMs as Semantic Encoders
BLaIR is a new benchmark and 570M-review dataset showing that LLM performance rankings on recommendation tasks have little correlation with rankings on general embedding benchmarks like MTEB.
-
Discovering Latent Knowledge in Language Models Without Supervision
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.
-
Rethinking the Role of Demonstrations: What Makes In-Context Learning Work?
Randomly replacing labels in in-context demonstrations barely hurts performance, showing that label space, input distribution, and sequence format drive in-context learning more than ground-truth labels.
-
SimCSE: Simple Contrastive Learning of Sentence Embeddings
SimCSE achieves 76.3% unsupervised and 81.6% supervised Spearman's correlation on STS tasks with BERT-base, improving prior best results by 4.2% and 2.2% via simple contrastive learning.
-
The Pile: An 800GB Dataset of Diverse Text for Language Modeling
The Pile is a newly constructed 825 GiB dataset from 22 diverse sources that enables language models to achieve better performance on academic, professional, and cross-domain tasks than models trained on Common Crawl variants.
-
Measuring Massive Multitask Language Understanding
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.
-
Language Models are Few-Shot Learners
GPT-3 shows that scaling an autoregressive language model to 175 billion parameters enables strong few-shot performance across diverse NLP tasks via in-context prompting without fine-tuning.
-
ELECTRA: Pre-training Text Encoders as Discriminators Rather Than Generators
ELECTRA replaces masked language modeling with replaced token detection, yielding contextual representations that outperform BERT at equal compute and match larger models like RoBERTa with far less compute.
-
REALM: Retrieval-Augmented Language Model Pre-Training
REALM augments language-model pre-training with an unsupervised retriever over Wikipedia documents and reports 4-16% absolute gains on open-domain QA benchmarks over prior implicit and explicit knowledge methods.
-
Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks
Sentence-BERT adapts BERT with siamese and triplet networks to produce sentence embeddings for efficient cosine-similarity comparisons, cutting computation time from hours to seconds on similarity search while matching BERT accuracy.
-
Less Effort, Shorter Proofs: Reinforcement Learning for Security Protocol Analysis in Tamarin
An RL-guided MCTS proof search for Tamarin finds more and shorter proofs than standard search across 16 protocol models.
-
Semantic Reranking at Inference Time for Hard Examples in Rhetorical Role Labeling
RISE is an inference-time semantic reranking framework that refines low-confidence predictions in rhetorical role labeling using contrastively learned label representations, delivering an average +9.15 macro-F1 gain on hard examples across eight datasets and seven models.
-
Scaling Laws from Sequential Feature Recovery: A Solvable Hierarchical Model
A solvable hierarchical model with power-law feature strengths yields explicit power-law scaling of prediction error through sequential recovery of latent directions by a layer-wise spectral algorithm.
-
BOOKMARKS: Efficient Active Storyline Memory for Role-playing
BOOKMARKS introduces searchable bookmarks as reusable answers to storyline questions, enabling active initialization and passive synchronization for more consistent role-playing agent memory than recurrent summarization.
-
Fin-Bias: Comprehensive Evaluation for LLM Decision-Making under human bias in Finance Domain
LLMs copy biased analyst ratings in investment decisions but a new detection method encourages independent reasoning and can improve stock return predictions beyond human levels.
-
Chain-based Distillation for Effective Initialization of Variable-Sized Small Language Models
Chain-based Distillation constructs a sequence of anchor models to enable efficient initialization of variable-sized SLMs through interpolation, with bridge distillation for cross-architecture transfer, yielding better performance than scratch training.
-
Is She Even Relevant? When BERT Ignores Explicit Gender Cues
A Dutch BERT model encodes gender linearly by epoch 20 but does not dynamically update its representations when explicit female cues contradict learned stereotypical associations in short sentence templates.
-
Beyond Factor Aggregation: Gauge-Aware Low-Rank Server Representations for Federated LoRA
GLoRA replaces raw factor averaging with gauge-aware aggregation in a consensus subspace estimated from client projectors, enabling consistent low-rank federated LoRA under heterogeneity.
-
Evaluating Non-English Developer Support in Machine Learning for Software Engineering
Code LLMs generate substantially worse comments outside English, and no tested automatic metric or LLM judge reliably matches human assessment of those outputs.
-
Adaptive Selection of LoRA Components in Privacy-Preserving Federated Learning
AS-LoRA adaptively chooses which LoRA factor to update per layer and round using a curvature-aware second-order score, eliminating reconstruction error floors and improving performance in DP federated learning.
-
Towards Self-Referential Analytic Assessment: A Profile-Based Approach to L2 Writing Evaluation with LLMs
LLMs outperform single human raters at spotting relative weaknesses in L2 writing profiles on the ICNALE GRA dataset while humans are better at spotting strengths, using a self-referential intra-learner evaluation method.
-
Leveraging Pretrained Language Models as Energy Functions for Glauber Dynamics Text Diffusion
Pretrained language models are used as energy functions for Glauber dynamics in discrete text diffusion, improving generation quality over prior diffusion LMs and matching autoregressive models on benchmarks and reasoning tasks.
-
Deep Graph-Language Fusion for Structure-Aware Code Generation
CGFuse enables deep token-level fusion of graph-derived structural features into language models, yielding 10-16% BLEU and 6-11% CodeBLEU gains on code generation tasks.
-
MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
MedStruct-S benchmark shows encoder-only models outperform larger decoder-only ones on key-conditioned QA from noisy OCR clinical reports, with fine-tuned large models winning only when scale is ignored.
-
How Language Models Process Negation
LLMs implement both attention-based suppression and constructive representations for negation, with construction dominant, despite poor accuracy from late-layer attention shortcuts.
-
Embedding-based In-Context Prompt Training for Enhancing LLMs as Text Encoders
EPIC trains LLMs to treat continuous embeddings as in-context prompts, yielding state-of-the-art text embedding performance on MTEB with or without prompts at inference and lower compute.
-
A Multi-View Media Profiling Suite: Resources, Evaluation, and Analysis
Presents MBFC-2025 dataset and multi-view embeddings with fusion methods for media bias and factuality, reporting SOTA results on ACL-2020 and new benchmarks on MBFC-2025.
-
InvEvolve: Evolving White-Box Inventory Policies via Large Language Models with Performance Guarantees
InvEvolve evolves white-box inventory policies from LLMs with statistical safety guarantees and outperforms classical and deep learning methods on synthetic and real retail data.
-
DEFault++: Automated Fault Detection, Categorization, and Diagnosis for Transformer Architectures
DEFault++ delivers automated hierarchical fault detection, categorization into 12 transformer-specific types, and root-cause diagnosis among 45 mechanisms on a new benchmark of 3,739 mutated instances, with AUROC >0.96 and Macro-F1 0.85, plus improved developer repair accuracy in a user study.
-
Auto-FlexSwitch: Efficient Dynamic Model Merging via Learnable Task Vector Compression
Auto-FlexSwitch achieves efficient dynamic model merging by decomposing task vectors into sparse masks, signs, and scalars, then making the compression learnable via gating and adaptive bit selection with KNN-based retrieval.
-
Agentic Adversarial Rewriting Exposes Architectural Vulnerabilities in Black-Box NLP Pipelines
A two-agent adversarial rewriting framework achieves 20-40% evasion rates against LLM-based misinformation detectors under strict black-box constraints with binary feedback only, far outperforming prior methods and linking success to specific architectural properties.
-
Not all ANIMALs are equal: metaphorical framing through source domains and semantic frames
An NLP framework shows that liberals and conservatives use different semantic frames within the same metaphorical source domains when discussing immigration, while also uncovering nuanced frames in climate change coverage.
-
A Mechanism and Optimization Study on the Impact of Information Density on User-Generated Content Named Entity Recognition
Low information density is identified as the root cause of NER failures on user-generated content, with the Window-Aware Optimization Module delivering up to 4.5% F1 gains and new SOTA on WNUT2017.
-
GuardPhish: Securing Open-Source LLMs from Phishing Abuse
Open-source LLMs detect phishing intent at high rates but still generate actionable phishing content, and GuardPhish supplies a dataset plus modular classifiers to close the gap.
-
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.
-
Psychological Steering of Large Language Models
Mean-difference residual stream injections outperform personality prompting for OCEAN trait steering in most LLMs, with hybrids performing best and showing approximate linearity but non-human trait covariances.
-
Human-Centric Topic Modeling with Goal-Prompted Contrastive Learning and Optimal Transport
GCTM-OT extracts goal candidates with an LLM, then uses goal-prompted contrastive learning and optimal transport to discover topics that are more coherent, diverse, and aligned with human intent than prior methods on subreddit data.
-
METRO: Towards Strategy Induction from Expert Dialogue Transcripts for Non-collaborative Dialogues
METRO induces both short-term actions and long-term planning from expert transcripts into a Strategy Forest, outperforming prior methods by 9-10% on two non-collaborative dialogue benchmarks.
-
A Hormone-inspired Emotion Layer for Transformer language models (HELT)
HormoneT5 augments T5 with a hormone-inspired block that predicts six continuous emotion values and uses them to modulate responses, reporting over 85% per-hormone accuracy and human preference for emotional quality.
-
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.
-
BMdataset: A Musicologically Curated LilyPond Dataset
A musicologically curated LilyPond dataset of 393 Baroque scores enables LilyBERT to outperform large-scale pre-training on composer and style classification when used alone for fine-tuning.
-
LASQ: A Low-resource Aspect-based Sentiment Quadruple Extraction Dataset
LASQ is a new quadruple extraction dataset for Uzbek and Uyghur that includes a syntax-aware model showing gains over baselines on the task.
-
Inductive Reasoning for Temporal Knowledge Graphs with Emerging Entities
TransFIR enables reasoning on temporal knowledge graphs for emerging entities by clustering them into semantic groups and borrowing interaction histories from similar known entities, yielding 28.6% average MRR gains.
-
Mask-Free Privacy Extraction and Rewriting: A Domain-Aware Approach via Prototype Learning
DAMPER learns domain privacy prototypes via contrastive learning and uses them to guide mask-free privacy extraction, preference-aligned rewriting, and differential privacy sampling for LLMs.
-
Attention Sink in Transformers: A Survey on Utilization, Interpretation, and Mitigation
The first survey on Attention Sink in Transformers structures the literature around fundamental utilization, mechanistic interpretation, and strategic mitigation.
-
Follow My Eyes: Backdoor Attacks on VLM-based Scanpath Prediction
Backdoor attacks on VLM-based scanpath predictors can redirect fixations toward chosen objects or inflate durations using input-conditioned triggers that evade cluster detection, and no tested defense blocks them without hurting clean accuracy.
-
Sell More, Play Less: Benchmarking LLM Realistic Selling Skill
SalesLLM provides an automatic evaluation framework for LLM sales dialogues that correlates 0.98 with human experts and shows top models approaching human performance while weaker ones lag.
-
iTAG: Inverse Design for Natural Text Generation with Accurate Causal Graph Annotations
iTAG generates natural text paired with accurate causal graph annotations by framing concept assignment as an inverse problem and refining selections via chain-of-thought reasoning until the text's relations align with the target causal structure.
-
Graph Topology Information Enhanced Heterogeneous Graph Representation Learning
ToGRL learns high-quality graph structures from raw heterogeneous graphs via a two-stage topology extraction process and prompt tuning, outperforming prior methods on five datasets.
-
The Indra Representation Hypothesis for Multimodal Alignment
Unimodal model representations converge to a relational structure captured by the Indra representation via V-enriched Yoneda embedding, which is unique and structure-preserving and improves cross-model and cross-modal robustness when instantiated with angular distance.