Depth-L transformers with W parameters have VC dimension Theta(L W log(T W)), yielding matching O(L W log((T+T')W)) upper and Omega(L W log((T+T')W/L)) lower bounds on sample complexity for chain-of-thought learning.
mega hub Canonical reference
GPT-4 Technical Report
Canonical reference. 76% of citing Pith papers cite this work as background.
2333
Pith papers citing it
Background
76% of classified citations
abstract
We report the development of GPT-4, a large-scale, multimodal model which can accept image and text inputs and produce text outputs. While less capable than humans in many real-world scenarios, GPT-4 exhibits human-level performance on various professional and academic benchmarks, including passing a simulated bar exam with a score around the top 10% of test takers. GPT-4 is a Transformer-based model pre-trained to predict the next token in a document. The post-training alignment process results in improved performance on measures of factuality and adherence to desired behavior. A core component of this project was developing infrastructure and optimization methods that behave predictably across a wide range of scales. This allowed us to accurately predict some aspects of GPT-4's performance based on models trained with no more than 1/1,000th the compute of GPT-4.
hub tools
citation-role summary
background 402
method 44
baseline 40
dataset 10
other 6
extension 1
citation-polarity summary
claims ledger
- abstract We report the development of GPT-4, a large-scale, multimodal model which can accept image and text inputs and produce text outputs. While less capable than humans in many real-world scenarios, GPT-4 exhibits human-level performance on various professional and academic benchmarks, including passing a simulated bar exam with a score around the top 10% of test takers. GPT-4 is a Transformer-based model pre-trained to predict the next token in a document. The post-training alignment process results in improved performance on measures of factuality and adherence to desired behavior. A core compone
authors
mega hub controls
export citing contexts JSON
export graph JSON
export full bundle JSON
open full Pith review
annotated reader queued
Recognition alignment
counterfactual ablation
co-cited works
representative citing papers
Across 30 LLMs and 205 TLA+ tasks, syntactic correctness reaches at most 26.6% and semantic correctness 8.6%, with all successes limited to progressive prompting and no advantage from larger models.
Sliding-window transformers without positional encodings are Turing complete because the sliding window breaks permutation symmetry and suffices to simulate Post machines via a constant-size histogram state.
Gaussian distributions are invariant under the mean-field Transformer flow, reducing infinite-dimensional dynamics to a bilinear control system on mean and covariance with explicit reachability and stability results.
TSFMAudit detects pretraining contamination in time series foundation models via probe adaptation dynamics (faster loss drop, smaller backbone shift), tested on 6 models and 187 datasets against 10 LLM-derived baselines.
NeuroQA is a large-scale 3D brain MRI visual question answering benchmark with verified image-grounded QA pairs, multi-domain coverage, and baseline evaluations showing current models lag behind text-only performance.
EgoIntrospect provides the first egocentric dataset with self-annotations for internal state tasks and shows multimodal LLMs struggle to infer subjective states from combined signals.
ViMU is the first benchmark for evaluating video models on metaphorical and subtextual understanding using hint-free questions grounded in multimodal evidence.
LongAct benchmark evaluates long-horizon household task execution from free-form instructions; HoloMind agent raises performance but top VLMs still reach only 59% goal completion and 16% full-task success.
CiteVQA requires models to cite specific document regions with bounding boxes alongside answers and finds that even the strongest MLLMs frequently cite the wrong region, with top SAA scores of only 76.0 for closed models and 22.5 for open-source ones.
LLM popularity judgments align more closely with pretraining data exposure counts than with Wikipedia popularity, with stronger effects in pairwise comparisons and larger models.
An MLLM-guided architecture with a mixture of frequency experts and relational alignment loss achieves state-of-the-art all-in-one image restoration, outperforming prior methods by up to 1.35 dB on the CDD11 dataset.
SignSGD provably beats SGD by a factor of d under sparse noise via matched ℓ1-norm upper and lower bounds, with an equivalent result for Muon on matrices, and this predicts faster GPT-2 pretraining.
IRIS-14B is the first LLM trained explicitly for GIMPLE-to-LLVM IR translation and outperforms much larger models by up to 44 percentage points on real-world C code.
ε-coresets for attention exist of size O(√d e^{ρ+o(ρ)}/ε) for unit-norm keys/values and queries of norm ≤ρ, nearly matching the Ω(√d e^ρ/ε) lower bound.
LLM agents achieve greater flexibility during execution by classifying actions via a reversibility taxonomy and using an Earliest-Conflict Rollback algorithm that matches full-restart quality while wasting far less completed work.
RespondeoQA is the first benchmark dataset for question answering and translation between Latin and English, with 7,800 pairs from pedagogical sources and initial LLM evaluations.
MathNet delivers the largest multilingual Olympiad math dataset and benchmarks where models like Gemini-3.1-Pro reach 78% on solving but embedding models struggle on equivalent problem retrieval, with retrieval augmentation yielding up to 12% gains.
ArgBench unifies 33 existing datasets into a standardized benchmark for testing LLMs across 46 argumentation tasks and analyzes the impact of prompting techniques and model factors on performance.
VoxSafeBench reveals that speech language models recognize social norms from text but fail to apply them when acoustic cues like speaker or scene determine the appropriate response.
RoboLab is a new simulation benchmark with 120 tasks across visual, procedural, and relational axes that quantifies generalization gaps and perturbation sensitivity in task-generalist robotic policies.
PhysInOne is a new dataset of 2 million videos across 153,810 dynamic 3D scenes covering 71 physical phenomena, shown to improve AI performance on physics-aware video generation, prediction, property estimation, and motion transfer.
HM-Bench is the first benchmark for MLLMs on hyperspectral images, showing models struggle with complex spatial-spectral reasoning and perform better with visual PCA images than textual reports.
ROTATE disentangles MLP neurons into faithful vocabulary channels by optimizing weight rotations to maximize vocabulary-space kurtosis, outperforming activation-based baselines for neuron descriptions.
citing papers explorer
-
Pretraining Exposure Explains Popularity Judgments in Large Language Models
LLM popularity judgments align more closely with pretraining data exposure counts than with Wikipedia popularity, with stronger effects in pairwise comparisons and larger models.
-
Leveraging Multimodal Large Language Models for All-in-One Image Restoration via a Mixture of Frequency Experts
An MLLM-guided architecture with a mixture of frequency experts and relational alignment loss achieves state-of-the-art all-in-one image restoration, outperforming prior methods by up to 1.35 dB on the CDD11 dataset.
-
When and Why SignSGD Outperforms SGD: A Theoretical Study Based on $\ell_1$-norm Lower Bounds
SignSGD provably beats SGD by a factor of d under sparse noise via matched ℓ1-norm upper and lower bounds, with an equivalent result for Muon on matrices, and this predicts faster GPT-2 pretraining.
-
ActivityForensics: A Comprehensive Benchmark for Localizing Manipulated Activity in Videos
ActivityForensics is the first large-scale benchmark for temporally localizing activity-level forgeries in videos, paired with a diffusion-based baseline called TADiff.
-
CANINE: Coaching Visually Impaired Users for Interactive Navigation with a Robot Guide Dog
CANINE decomposes interactive navigation into sub-skills, prioritizes training via knowledge tracing, and uses foundation models for adaptive error correction, yielding better learning efficiency and navigation performance than generic instructions in user studies.
-
LEAF-SQL: Level-wise Exploration with Adaptive Fine-graining for Text-to-SQL Skeleton Prediction
LEAF-SQL uses level-wise exploration with adaptive fine-graining and dual agents to generate diverse SQL skeletons, reaching 71.6% execution accuracy on the BIRD benchmark and outperforming prior search- and skeleton-based methods.
-
Text-to-CAD Evaluation with CADTests
Introduces CADTestBench as a test-based benchmark for Text-to-CAD and shows that using CADTests to guide generation produces simple baselines outperforming prior methods.
-
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.
-
MuSS: A Large-Scale Dataset and Cinematic Narrative Benchmark for Multi-Shot Subject-to-Video Generation
MuSS is a new movie-sourced dataset and benchmark that enables AI models to generate multi-shot videos with improved narrative coherence and subject identity preservation.
-
Character Beyond Speech: Leveraging Role-Playing Evaluation in Audio Large Language Models via Reinforcement Learning
RoleJudge is a multidimensional evaluation framework for speech-character alignment in audio LLMs, backed by the RoleChat dataset and multi-stage RL training with standard alignment to reduce reward issues.
-
WildDet3D: Scaling Promptable 3D Detection in the Wild
WildDet3D is a promptable 3D detector paired with a new 1M-image dataset across 13.5K categories that sets SOTA on open-world and zero-shot 3D detection benchmarks.
-
PRIME: Training Free Proactive Reasoning via Iterative Memory Evolution for User-Centric Agent
PRIME enables agents to proactively reason in user-centric tasks by iteratively evolving structured memories from interaction trajectories without gradient-based training.
-
Q-Mask: Query-driven Causal Masks for Text Anchoring in OCR-Oriented Vision-Language Models
Q-Mask uses query-conditioned causal masks to separate text location from recognition in OCR VLMs, backed by a new benchmark and 26M-pair training dataset.
-
OpenClassGen: A Large-Scale Corpus of Real-World Python Classes for LLM Research
OpenClassGen supplies 324,843 real-world Python classes with self-contained skeletons and static metrics to support LLM class generation research and evaluation.
-
Structured 3D Latents for Scalable and Versatile 3D Generation
SLAT provides a unified 3D latent representation enabling versatile high-quality generation across multiple output formats from text or image inputs.
-
Voyager: An Open-Ended Embodied Agent with Large Language Models
Voyager achieves superior lifelong learning in Minecraft by combining an automatic exploration curriculum, a library of executable skills, and iterative LLM prompting with environment feedback, yielding 3.3x more unique items and 15.3x faster milestone unlocks than prior methods while generalizing技能
-
Distinguishable Deletion: Unifying Knowledge Erasure and Refusal for Large Language Model Unlearning
Distinguishable Deletion unifies knowledge erasure and refusal for LLM unlearning via an energy index that enforces boundaries during training and enables refusal at inference.
-
Scalable Token-Level Hallucination Detection in Large Language Models
TokenHD uses a scalable data synthesis engine and importance-weighted training to create token-level hallucination detectors that work on free-form text and scale from 0.6B to 8B parameters, outperforming larger reasoning models.
-
Explaining and Breaking the Safety-Helpfulness Ceiling via Preference Dimensional Expansion
MORA breaks the safety-helpfulness ceiling in LLMs by pre-sampling single-reward prompts and rewriting them to incorporate multi-dimensional intents, delivering 5-12.4% gains in sequential alignment and 4.6% overall improvement in simultaneous alignment.
-
LimeCross: Context-Conditioned Layered Image Editing with Structural Consistency
LimeCross enables text-guided editing of individual layers in composite images by conditioning on cross-layer context via bi-stream attention while preserving layer integrity and introducing the LayerEditBench benchmark.
-
ChartREG++: Towards Benchmarking and Improving Chart Referring Expression Grounding under Diverse referring clues and Multi-Target Referring
ChartREG++ benchmark for multi-target chart referring expression grounding with diverse clues plus a plotting-code synthesis pipeline for accurate masks that improves model performance.
-
Bridging the Embodiment Gap: Disentangled Cross-Embodiment Video Editing
A dual-contrastive disentanglement method factorizes videos into independent task and embodiment latents, then uses a parameter-efficient adapter on a frozen video diffusion model to synthesize robot executions from single human demonstrations without paired data.
-
Embody4D: A Generalist Data Engine for Embodied 4D World Modeling
Embody4D generates novel-view videos from monocular robot videos via a 3D-aware synthesis pipeline, confidence-aware expert modulation, and interaction-aware attention for embodied 4D world modeling.
-
Affordance Agent Harness: Verification-Gated Skill Orchestration
Affordance Agent Harness is a verification-gated orchestration system that unifies skills via an evidence store, episodic memory priors, an adaptive router, and a self-consistency verifier to improve accuracy-cost tradeoffs in open-world affordance grounding.
-
DASH-KV: Accelerating Long-Context LLM Inference via Asymmetric KV Cache Hashing
DASH-KV accelerates long-context LLM inference to linear complexity via asymmetric KV cache hashing and mixed-precision retention, matching full attention performance on LongBench.
-
On the Importance and Evaluation of Narrativity in Natural Language AI Explanations
XAI explanations should be narratives with continuous structure, cause-effect, fluency and diversity, and new metrics are needed to evaluate this better than standard NLP scores.
-
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.
-
Adaptive Test-Time Compute Allocation for Reasoning LLMs via Constrained Policy Optimization
A Lagrangian-relaxation plus imitation-learning pipeline adaptively allocates test-time compute to LLMs, outperforming uniform baselines by up to 12.8% relative accuracy on MATH while staying within a fixed average budget.
-
Label Effects: Shared Heuristic Reliance in Trust Assessment by Humans and LLM-as-a-Judge
Both humans and LLMs trust content more when labeled human-authored than AI-generated, with LLMs showing denser attention to labels and higher uncertainty under AI labels, mirroring human heuristic patterns.
-
SpatialEdit: Benchmarking Fine-Grained Image Spatial Editing
SpatialEdit provides a benchmark, large synthetic dataset, and baseline model for precise object and camera spatial manipulations in images, with the model beating priors on spatial editing.
-
InsTraj: Instructing Diffusion Models with Travel Intentions to Generate Real-world Trajectories
InsTraj generates realistic, instruction-faithful GPS trajectories by using an LLM to parse natural-language travel intent and a multimodal diffusion transformer to produce the paths.
-
CoopGuard: Stateful Cooperative Agents Safeguarding LLMs Against Evolving Multi-Round Attacks
CoopGuard deploys cooperative agents to track conversation history and counter evolving multi-round attacks on LLMs, achieving a 78.9% reduction in attack success rate on a new 5,200-sample benchmark.
-
EvaNet: Towards More Efficient and Consistent Infrared and Visible Image Fusion Assessment
EvaNet is a lightweight network that efficiently approximates image fusion metrics with improved consistency to human perception via decomposition, contrastive learning, and LLM input.
-
Hierarchical, Interpretable, Label-Free Concept Bottleneck Model
HIL-CBM is a hierarchical label-free concept bottleneck model that improves classification accuracy and explanation quality over prior single-level CBMs using a visual consistency loss and dual heads.
-
Robotic Manipulation by Imitating Generated Videos Without Physical Demonstrations
RIGVid shows that filtered AI-generated videos can serve as effective supervision for complex robotic manipulation tasks without any real demonstrations.
-
ImgEdit: A Unified Image Editing Dataset and Benchmark
ImgEdit supplies 1.2 million curated edit pairs and a three-part benchmark that let a VLM-based model outperform prior open-source editors on adherence, quality, and detail preservation.
-
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.
-
RAPTOR: Recursive Abstractive Processing for Tree-Organized Retrieval
RAPTOR introduces a tree-organized retrieval method using recursive abstractive summaries, achieving a 20% absolute accuracy improvement on the QuALITY benchmark when paired with GPT-4.
-
Generating Place-Based Compromises Between Two Points of View
Empathic similarity feedback in prompts generates more acceptable compromises than chain-of-thought, and margin-based training on the resulting data lets smaller models produce them without ongoing empathy estimation.
-
From Coarse to Fine: Self-Adaptive Hierarchical Planning for LLM Agents
AdaPlan-H enables LLM agents to generate self-adaptive hierarchical plans that adjust detail level to task difficulty, improving success rates in multi-step tasks.
-
A Graph-Enhanced Defense Framework for Explainable Fake News Detection with LLM
G-Defense builds claim-centered graphs from sub-claims, applies RAG for evidence and competing explanations, then uses graph inference to detect fake news veracity and generate intuitive explanation graphs, claiming SOTA results.
-
COIVis: Eye-tracking-based Visual Exploration of Concept Learning in MOOC Videos
COIVis aligns multimodal video concepts with screen space and time to turn eye-tracking data into interpretable learner-state sequences, enabling instructors to explore cohort and individual learning patterns in MOOCs.
-
Aligning Modalities in Vision Large Language Models via Preference Fine-tuning
POVID generates AI-created preference data to fine-tune vision-language models with DPO, reducing hallucinations and improving benchmark scores.
-
Network Edge Inference for Large Language Models: Principles, Techniques, and Opportunities
A survey synthesizing challenges, system architectures, model optimizations, deployment methods, and resource management techniques for large language model inference at the network edge.