WildChat releases a dataset of 1 million ChatGPT conversations with timestamps, demographics, and headers, claimed to be the most diverse and multilingual such resource available.
Title resolution pending
52 Pith papers cite this work. Polarity classification is still indexing.
52
Pith papers citing it
citation-role summary
background 3
citation-polarity summary
roles
background 2representative citing papers
Pre-pretraining on MP-STRUCT matches k-Shuffle Dyck baselines in efficiency while adding human-like resistance to implausible languages and challenges the need for C-RASP definability in effective PPT languages.
QAOD projects away question-aligned directions from answer representations to isolate domain-agnostic factuality signals, enabling efficient hallucination detection with top in-domain AUROC and up to 21% better OOD transfer.
Energy-navigated trajectory shaping during training produces 8-step discrete flow matching students that achieve 32% lower perplexity than 1024-step teachers on 170M language models with unchanged inference cost.
OrchJail uses orchestration-guided fuzzing to jailbreak tool-calling T2I agents by targeting high-risk tool patterns, yielding higher attack success rates, better image quality, and lower costs than prior prompt-only methods.
New datasets and language model classifiers enable detection of misinformation spans in video transcripts with an F1 score of 0.68.
SWE-RL uses RL on software evolution data to train LLMs achieving 41% on SWE-bench Verified with generalization to other reasoning tasks.
Absorbing discrete diffusion models the conditional distributions of clean data; reparameterizing yields a time-independent RADD that unifies with AO-ARMs and reaches SOTA perplexity among diffusion models on zero-shot language benchmarks.
Gated SAEs decouple which features to use from how large their activations should be, applying the L1 penalty only to selection and thereby eliminating shrinkage while halving the number of firing features needed for good fidelity.
RouterBench supplies a standardized benchmark, 405k+ inference dataset, theoretical framework, and comparative analysis for multi-LLM routing systems.
HSTU-based generative recommenders with 1.5 trillion parameters scale as a power law with compute up to GPT-3 scale, outperform baselines by up to 65.8% NDCG, run 5-15x faster than FlashAttention2 on long sequences, and improve online A/B metrics by 12.4%.
MultiHop-RAG is a new benchmark dataset demonstrating that existing retrieval-augmented generation systems perform poorly on multi-hop queries requiring retrieval and reasoning over multiple evidence pieces.
Q-Align trains LMMs on discrete text-defined levels for visual scoring, achieving SOTA on IQA, IAA, and VQA while unifying the tasks in OneAlign.
LoRA adapters should be scaled by 1/sqrt(rank) rather than 1/rank to stabilize learning and enable effective use of higher ranks during fine-tuning of large language models.
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.
Min-K% Prob detects pretraining data in LLMs by flagging outlier low-probability words in text, achieving 7.4% better performance than prior methods on the new WIKIMIA benchmark.
Introduces predictive prefetching for RAG that anticipates retrieval needs several tokens ahead via three components, reporting up to 43.5% latency reduction and 62.4% TTFT improvement while preserving answer quality.
ATD-Trans is a new geographically annotated Japanese-English travelogue dataset that reveals Japanese-enhanced models perform better on geo-entity translation while domestic Japanese locations remain harder to translate accurately.
Risk-sensitive preference games using convex risk measures produce policies that are robust across data strata and match or exceed standard Nash learning performance without added cost.
Structured Recurrent Mixers enable algebraic switching between parallel training and recurrent inference representations, yielding higher throughput, concurrency, and training efficiency than comparable linear-complexity models on language tasks.
ExecuTorch is a unified PyTorch-native deployment framework that enables seamless on-device execution of AI models across heterogeneous hardware while preserving original PyTorch semantics.
PUPPET jointly optimizes LLM outputs for high detectability and task performance via RL rewards from a detector and a task evaluator, outperforming watermarking on tasks while matching detectability.
GRASPrune removes 50% of parameters from LLaMA-2-7B via global gating and projected straight-through estimation, reaching 12.18 WikiText-2 perplexity and competitive zero-shot accuracy after four epochs on 512 calibration sequences.
LIMO achieves 63.3% on AIME24 and 95.6% on MATH500 via supervised fine-tuning on roughly 1% of the data used by prior models, supporting the claim that minimal strategic examples suffice when pre-training has already encoded domain knowledge.
citing papers explorer
-
Language Acquisition Device in Large Language Models
Pre-pretraining on MP-STRUCT matches k-Shuffle Dyck baselines in efficiency while adding human-like resistance to implausible languages and challenges the need for C-RASP definability in effective PPT languages.
-
When Answers Stray from Questions: Hallucination Detection via Question-Answer Orthogonal Decomposition
QAOD projects away question-aligned directions from answer representations to isolate domain-agnostic factuality signals, enabling efficient hallucination detection with top in-domain AUROC and up to 21% better OOD transfer.
-
Trajectory as the Teacher: Few-Step Discrete Flow Matching via Energy-Navigated Distillation
Energy-navigated trajectory shaping during training produces 8-step discrete flow matching students that achieve 32% lower perplexity than 1024-step teachers on 170M language models with unchanged inference cost.
-
OrchJail: Jailbreaking Tool-Calling Text-to-Image Agents by Orchestration-Guided Fuzzing
OrchJail uses orchestration-guided fuzzing to jailbreak tool-calling T2I agents by targeting high-risk tool patterns, yielding higher attack success rates, better image quality, and lower costs than prior prompt-only methods.
-
Misinformation Span Detection in Videos via Audio Transcripts
New datasets and language model classifiers enable detection of misinformation spans in video transcripts with an F1 score of 0.68.
-
SWE-RL: Advancing LLM Reasoning via Reinforcement Learning on Open Software Evolution
SWE-RL uses RL on software evolution data to train LLMs achieving 41% on SWE-bench Verified with generalization to other reasoning tasks.
-
Improving Dictionary Learning with Gated Sparse Autoencoders
Gated SAEs decouple which features to use from how large their activations should be, applying the L1 penalty only to selection and thereby eliminating shrinkage while halving the number of firing features needed for good fidelity.
-
RouterBench: A Benchmark for Multi-LLM Routing System
RouterBench supplies a standardized benchmark, 405k+ inference dataset, theoretical framework, and comparative analysis for multi-LLM routing systems.
-
Actions Speak Louder than Words: Trillion-Parameter Sequential Transducers for Generative Recommendations
HSTU-based generative recommenders with 1.5 trillion parameters scale as a power law with compute up to GPT-3 scale, outperform baselines by up to 65.8% NDCG, run 5-15x faster than FlashAttention2 on long sequences, and improve online A/B metrics by 12.4%.
-
A Rank Stabilization Scaling Factor for Fine-Tuning with LoRA
LoRA adapters should be scaled by 1/sqrt(rank) rather than 1/rank to stabilize learning and enable effective use of higher ranks during fine-tuning of large language models.
-
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.
-
Predictive Prefetching for Retrieval-Augmented Generation
Introduces predictive prefetching for RAG that anticipates retrieval needs several tokens ahead via three components, reporting up to 43.5% latency reduction and 62.4% TTFT improvement while preserving answer quality.
-
Structure from Strategic Interaction & Uncertainty: Risk Sensitive Games for Robust Preference Learning
Risk-sensitive preference games using convex risk measures produce policies that are robust across data strata and match or exceed standard Nash learning performance without added cost.
-
ExecuTorch -- A Unified PyTorch Solution to Run AI Models On-Device
ExecuTorch is a unified PyTorch-native deployment framework that enables seamless on-device execution of AI models across heterogeneous hardware while preserving original PyTorch semantics.
-
LLM Output Detectability and Task Performance Can be Jointly Optimized
PUPPET jointly optimizes LLM outputs for high detectability and task performance via RL rewards from a detector and a task evaluator, outperforming watermarking on tasks while matching detectability.
-
GRASPrune: Global Gating for Budgeted Structured Pruning of Large Language Models
GRASPrune removes 50% of parameters from LLaMA-2-7B via global gating and projected straight-through estimation, reaching 12.18 WikiText-2 perplexity and competitive zero-shot accuracy after four epochs on 512 calibration sequences.
-
LIMO: Less is More for Reasoning
LIMO achieves 63.3% on AIME24 and 95.6% on MATH500 via supervised fine-tuning on roughly 1% of the data used by prior models, supporting the claim that minimal strategic examples suffice when pre-training has already encoded domain knowledge.
-
Agent Q: Advanced Reasoning and Learning for Autonomous AI Agents
Agent Q integrates MCTS-guided search, self-critique, and off-policy DPO to train LLM agents that outperform behavior cloning and reinforced fine-tuning baselines in WebShop and achieve up to 95.4% success in real-world booking scenarios.
-
ALLaVA: Harnessing GPT4V-Synthesized Data for Lite Vision-Language Models
ALLaVA creates 1.3M GPT4V-synthesized samples enabling 4B VLMs to achieve competitive results on 17 benchmarks and match 7B/13B models on some tasks.
-
Steering Llama 2 via Contrastive Activation Addition
Contrastive Activation Addition steers Llama 2 Chat by adding averaged residual-stream activation differences from contrastive example pairs to control targeted behaviors at inference time.
-
The Geometry of Truth: Emergent Linear Structure in Large Language Model Representations of True/False Datasets
At sufficient scale, LLMs linearly represent the truth value of factual statements, as shown by visualizations, cross-dataset generalization, and causal interventions that flip truth judgments.
-
Chain-of-Verification Reduces Hallucination in Large Language Models
Chain-of-Verification reduces hallucinations in large language models by drafting responses, planning independent verification questions, answering them separately, and generating a final verified output.
-
Scaling Relationship on Learning Mathematical Reasoning with Large Language Models
Pre-training loss predicts LLM math reasoning better than parameter count; rejection sampling fine-tuning with diverse paths raises LLaMA-7B accuracy on GSM8K from 35.9% with SFT to 49.3%.
-
The Efficiency Gap in Byte Modeling
Byte modeling incurs greater scaling overhead for masked diffusion than autoregressive models because the diffusion objective destroys local byte contiguity needed to resolve semantics.
-
From Flat Facts to Sharp Hallucinations: Detecting Stubborn Errors via Gradient Sensitivity
EPGS detects high-confidence factual errors in LLMs by using embedding perturbations to measure gradient sensitivity as a proxy for sharp versus flat minima.
-
Can Continual Pre-training Bridge the Performance Gap between General-purpose and Specialized Language Models in the Medical Domain?
Continual pre-training on a German medical corpus lets 7B models close much of the performance gap with 24B general models on medical benchmarks, though merging introduces some language mixing and verbosity.
-
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.
-
MASS-RAG: Multi-Agent Synthesis Retrieval-Augmented Generation
MASS-RAG uses distinct agents for evidence summarization, extraction, and reasoning, then synthesizes their outputs to improve answer quality over standard RAG baselines on four benchmarks, especially when evidence is distributed.
-
AppAgent: Multimodal Agents as Smartphone Users
AppAgent lets large language models operate diverse smartphone apps via visual interactions and learns app usage from exploration or demonstrations.
-
Detecting Language Model Attacks with Perplexity
Jailbreak prompts with adversarial suffixes have high GPT-2 perplexity, and a LightGBM model on perplexity and length detects most attacks.