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.
hub
Can a suit of armor conduct electricity? a new dataset for open book question answering
44 Pith papers cite this work, alongside 635 external citations. Polarity classification is still indexing.
44
Pith papers citing it
635
external citations · OpenAlex
hub tools
citation-role summary
background 2
method 1
other 1
citation-polarity summary
representative citing papers
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.
UniRank introduces dual local-global scoring for rank allocation in LLM low-rank decomposition plus rank-preserving fine-tuning, achieving up to 50% lower perplexity than uniform baselines in one-shot compression.
UrduMMLU is a new native-source MCQ benchmark for Urdu that reveals top LLMs reach only ~90% accuracy with large gaps on region-specific humanities content.
TokAlign++ learns token alignments between LLM vocabularies from monolingual representations to enable faster adaptation, better text compression, and effective token-level distillation across 15 languages with minimal steps.
ConQuR is a post-training rotation calibration technique that aligns activations to hypercube corners via Procrustes optimization and online updates, delivering competitive LLM quantization performance without end-to-end training or offline activation storage.
Gaussian Kernel Attention replaces learned QKV projections with a Gaussian RBF kernel on per-head token features, using 0.42x parameters and 0.49x FLOPs while showing competitive language modeling performance at depth 20.
GAIA benchmark shows humans at 92% accuracy on simple real-world questions far outperform current AI systems at 15%, proposing this gap as a key milestone for general AI.
A stateless, SVD-free regularizer approximates polar factors to induce low-rank weight structure during training, enabling better post-training compression of vision models and LLMs at under 8% overhead.
Pruning MoE LLMs according to the stationary distribution of a Markov chain over (layer, expert) routing transitions retains more task performance than local importance heuristics, with up to ~3.5% relative gains over the strongest baseline at 50% compression.
Optimal layer-patching order for boomerang distillation is a shortest path on a KL-weighted Boolean lattice; greedy KLPatch and simple sequential orders often yield near-optimal interpolations.
SCAPE enables 90-99% sparse gradient communication in sharded Adam-style LLM training by deriving masks from first-moment statistics, achieving up to 43.3% faster pre-training on Llama-500M with no loss in validation loss or downstream accuracy.
Pruning attention layers in five LLMs across eight datasets maintains accuracy but degrades faithfulness and calibration.
Safety Reflection Pretraining adds regular safety reflections to pretraining data to integrate self-monitoring and reduce unsafe generalization from safe data in LLMs.
MLP activations measured as massive activations or first four moments correlate weakly (max |Spearman| = 0.33) with in-context example quality across Llama-3.2-3B, Qwen2.5-3B, and multiple classification/generative tasks, so activation-based active learning should not be used for ICL.
SSMoE uses eigenvectors of expert weights via SVD to build training-free non-collapsing routers for SMoE models in language and vision tasks.
NaRA conditions LoRA updates on noise level via a shared hypernetwork for diffusion LLMs, claiming better performance than static LoRA on reasoning and code tasks with negligible overhead.
Proxy metrics from next-token distributions over expert solutions outperform loss and compute baselines for ranking LLMs, selecting pretraining data, and extrapolating performance across compute scales.
SP-KV trains a utility predictor jointly with the LLM to dynamically prune low-utility KV cache entries, achieving 3-10x memory reduction during generation with negligible performance loss.
Training transformers with KV sparsification during continued pretraining produces representations that admit better post-hoc KV cache compression, improving quality under memory budgets for long-context tasks.
MOONSHOT is a wrapper framework that combines reconstruction loss and Taylor approximation objectives to improve one-shot pruning results on LLMs and vision models.
Layer-wise LLM rotations can be fused offline and residual basis mismatch approximated by a low-rank subspace, matching expensive layer-wise PTQ accuracy with near-global overhead.
EdgeFlow reduces mobile LLM cold-start latency up to 4.07x versus llama.cpp, MNN, and llm.npu by NPU-aware adaptive quantization, SIMD-friendly packing, and synergistic granular CPU-NPU pipelining at comparable accuracy.
TalkLoRA equips MoE-LoRA experts with a communication module that smooths routing dynamics and improves performance on language tasks under similar parameter budgets.
citing papers explorer
-
Faithfulness Metrics Don't Measure Faithfulness: A Meta-Evaluation with Ground Truth
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.
-
MultiHashFormer: Hash-based Generative Language Models
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.
-
UniRank: Unified Rank Allocation for Low-Rank LLM Compression
UniRank introduces dual local-global scoring for rank allocation in LLM low-rank decomposition plus rank-preserving fine-tuning, achieving up to 50% lower perplexity than uniform baselines in one-shot compression.
-
UrduMMLU: A Massive Multitask Benchmark for Urdu Language Understanding
UrduMMLU is a new native-source MCQ benchmark for Urdu that reveals top LLMs reach only ~90% accuracy with large gaps on region-specific humanities content.
-
TokAlign++: Advancing Vocabulary Adaptation via Better Token Alignment
TokAlign++ learns token alignments between LLM vocabularies from monolingual representations to enable faster adaptation, better text compression, and effective token-level distillation across 15 languages with minimal steps.
-
ConQuR: Corner Aligned Activation Quantization via Optimized Rotations for LLMs
ConQuR is a post-training rotation calibration technique that aligns activations to hypercube corners via Procrustes optimization and online updates, delivering competitive LLM quantization performance without end-to-end training or offline activation storage.
-
Projection-Free Transformers via Gaussian Kernel Attention
Gaussian Kernel Attention replaces learned QKV projections with a Gaussian RBF kernel on per-head token features, using 0.42x parameters and 0.49x FLOPs while showing competitive language modeling performance at depth 20.
-
GAIA: a benchmark for General AI Assistants
GAIA benchmark shows humans at 92% accuracy on simple real-world questions far outperform current AI systems at 15%, proposing this gap as a key milestone for general AI.
-
SLORR: Simple and Efficient In-Training Low-Rank Regularization
A stateless, SVD-free regularizer approximates polar factors to induce low-rank weight structure during training, enabling better post-training compression of vision models and LLMs at under 8% overhead.
-
It Takes a MAESTRO To Prune Bad Experts
Pruning MoE LLMs according to the stationary distribution of a Markov chain over (layer, expert) routing transitions retains more task performance than local importance heuristics, with up to ~3.5% relative gains over the strongest baseline at 50% compression.
-
Understanding Layer Patching in Model Size Interpolation
Optimal layer-patching order for boomerang distillation is a shortest path on a KL-weighted Boolean lattice; greedy KLPatch and simple sequential orders often yield near-optimal interpolations.
-
SCAPE: Accurate and Efficient LLM Training with Extreme Sparse Communication
SCAPE enables 90-99% sparse gradient communication in sharded Adam-style LLM training by deriving masks from first-moment statistics, achieving up to 43.3% faster pre-training on Llama-500M with no loss in validation loss or downstream accuracy.
-
Don't Go Breaking My LLM: The Impact of Pruning Attention Layers on Explanation Faithfulness and Confidence Calibration
Pruning attention layers in five LLMs across eight datasets maintains accuracy but degrades faithfulness and calibration.
-
Beyond Safe Data: Pretraining-Stage Alignment with Regular Safety Reflection
Safety Reflection Pretraining adds regular safety reflections to pretraining data to integrate self-monitoring and reduce unsafe generalization from safe data in LLMs.
-
Activation-Based Active Learning for In-Context Learning: Challenges and Insights
MLP activations measured as massive activations or first four moments correlate weakly (max |Spearman| = 0.33) with in-context example quality across Llama-3.2-3B, Qwen2.5-3B, and multiple classification/generative tasks, so activation-based active learning should not be used for ICL.
-
Eigenvectors of Experts are Training-free Non-collapsing Routers
SSMoE uses eigenvectors of expert weights via SVD to build training-free non-collapsing routers for SMoE models in language and vision tasks.
-
NaRA: Noise-Aware LoRA for Parameter-Efficient Fine-Tuning of Diffusion LLMs
NaRA conditions LoRA updates on noise level via a shared hypernetwork for diffusion LLMs, claiming better performance than static LoRA on reasoning and code tasks with negligible overhead.
-
Forecasting Downstream Performance of LLMs With Proxy Metrics
Proxy metrics from next-token distributions over expert solutions outperform loss and compute baselines for ranking LLMs, selecting pretraining data, and extrapolating performance across compute scales.
-
Self-Pruned Key-Value Attention: Learning When to Write by Predicting Future Utility
SP-KV trains a utility predictor jointly with the LLM to dynamically prune low-utility KV cache entries, achieving 3-10x memory reduction during generation with negligible performance loss.
-
Training Transformers for KV Cache Compressibility
Training transformers with KV sparsification during continued pretraining produces representations that admit better post-hoc KV cache compression, improving quality under memory budgets for long-context tasks.
-
MOONSHOT : A Framework for Multi-Objective Pruning of Vision and Large Language Models
MOONSHOT is a wrapper framework that combines reconstruction loss and Taylor approximation objectives to improve one-shot pruning results on LLMs and vision models.
-
ReSpinQuant: Efficient Layer-Wise LLM Quantization via Subspace Residual Rotation Approximation
Layer-wise LLM rotations can be fused offline and residual basis mismatch approximated by a low-rank subspace, matching expensive layer-wise PTQ accuracy with near-global overhead.
-
EdgeFlow: Fast Cold Starts for LLMs on Mobile Devices
EdgeFlow reduces mobile LLM cold-start latency up to 4.07x versus llama.cpp, MNN, and llm.npu by NPU-aware adaptive quantization, SIMD-friendly packing, and synergistic granular CPU-NPU pipelining at comparable accuracy.
-
TalkLoRA: Communication-Aware Mixture of Low-Rank Adaptation for Large Language Models
TalkLoRA equips MoE-LoRA experts with a communication module that smooths routing dynamics and improves performance on language tasks under similar parameter budgets.
-
Does Math Reasoning Improve General LLM Capabilities? Understanding Transferability of LLM Reasoning
Math reasoning gains in LLMs rarely transfer to general domains; RL tuning generalizes while SFT causes forgetting and representation drift.
-
Condense, Don't Just Prune: Enhancing Efficiency and Performance in MoE Layer Pruning
CD-MoE condenses fine-grained MoE layers with shared experts into dense layers, retaining 90% accuracy with 27.5% memory cut and 1.26x speedup on DeepSeekMoE-16B, recovering 98% via brief fine-tuning.
-
DataComp-LM: In search of the next generation of training sets for language models
DCLM-Baseline dataset lets a 7B model reach 64% 5-shot MMLU accuracy after 2.6T tokens, beating prior open-data models by 6.6 points on MMLU with 40% less compute.
-
Gemini: A Family of Highly Capable Multimodal Models
Gemini Ultra reaches human-expert performance on MMLU for the first time and sets new state-of-the-art results on 30 of 32 benchmarks, including all 20 multimodal ones tested.
-
Enhancing Chat Language Models by Scaling High-quality Instructional Conversations
UltraChat supplies 1.5 million high-quality multi-turn dialogues that, when used to fine-tune LLaMA, produce UltraLLaMA, which outperforms prior open-source chat models including Vicuna.
-
BloombergGPT: A Large Language Model for Finance
BloombergGPT is a 50B parameter LLM trained on a 708B token mixed financial and general dataset that outperforms prior models on financial benchmarks while preserving general LLM performance.
-
Atlas: Few-shot Learning with Retrieval Augmented Language Models
Atlas reaches over 42% accuracy on Natural Questions with only 64 examples, outperforming a 540B-parameter model by 3% with 50x fewer parameters.
-
GPT-NeoX-20B: An Open-Source Autoregressive Language Model
GPT-NeoX-20B is a publicly released 20B parameter autoregressive language model trained on the Pile that shows strong gains in five-shot reasoning over similarly sized prior models.
-
Matching Tasks to Objectives: Fine-Tuning and Prompt-Tuning Strategies for Encoder-Decoder Pre-trained Language Models
Introduces MTO framework for matching tasks to pre-training objectives in encoder-decoder models, achieving over 120% performance gains in few-shot commonsense tasks.
-
GRASP: Geometry-aware Residual Alignment for Scalable Pretraining Data Attribution
GRASP is a scalable method for subset-level data attribution in pretraining that models interactions via a geometry-aware quadratic penalty and claims to double rank correlation while cutting costs.
-
PC Layer: Polynomial Weight Preconditioning for Improving LLM Pre-Training
A polynomial preconditioning layer controls singular value spectra of transformer weights to stabilize pre-training, shown effective on Llama-1B and supported by convergence theory for deep linear networks.
-
What Should Agents Say? Action-state Communication for Efficient Multi-Agent Systems
Introduces PACT protocol that projects agent outputs into action-state records, yielding comparable or better task performance with substantially fewer tokens in multi-agent LLM systems and production harnesses.
-
IO-SVD: Input-Output Whitened SVD for Adaptive-Rank LLM Compression
IO-SVD performs SVD-based LLM compression by constructing a KL-aware double-sided whitening space and using first-order loss estimates for heterogeneous rank allocation.
-
Gated Delta Networks: Improving Mamba2 with Delta Rule
Gated DeltaNet integrates gating and delta rules into linear transformers, outperforming Mamba2 and DeltaNet on language modeling, reasoning, retrieval, and long-context tasks.
-
InternLM2 Technical Report
InternLM2 is a new open-source LLM that outperforms prior versions on 30 benchmarks and long-context tasks through scaled pre-training to 32k tokens and a conditional online RLHF alignment strategy.
-
PaLM 2 Technical Report
PaLM 2 reports state-of-the-art results on language, reasoning, and multilingual tasks with improved efficiency over PaLM.
-
Ling and Ring 2.6 Technical Report: Efficient and Instant Agentic Intelligence at Trillion-Parameter Scale
Technical report announcing Ling-2.6 and Ring-2.6 models with hybrid linear attention, evolutionary CoT, and KPop RL for efficient agentic intelligence at scale.
-
Benchmarked Yet Not Measured -- Generative AI Should be Evaluated Against Real-World Utility
Generative AI evaluation must shift from static benchmark scores to measuring sustained improvements in human capabilities within specific deployment contexts.
-
Learning in the Fisher Subspace: A Guided Initialization for LoRA Fine-Tuning
Fisher information from the target data distribution supplies a task-dependent criterion for selecting LoRA directions that outperforms weight-magnitude heuristics.
-
A Tree-of-Thoughts Inspired Hybrid Approach for Legal Case Judgement Summarization using LLMs
A tree-of-thoughts inspired hybrid extractive-abstractive LLM prompt yields better legal case judgment summaries than standard extractive or abstractive prompts.