MMMU provides 11.5K heterogeneous college-level multimodal questions that current models solve at 56-59% accuracy, establishing a new standard for expert multimodal evaluation.
super hub Canonical reference
PaLM: Scaling Language Modeling with Pathways
Canonical reference. 83% of citing Pith papers cite this work as background.
198
Pith papers citing it
2,131
external citations · Pith
Background
83% of classified citations
abstract
Large language models have been shown to achieve remarkable performance across a variety of natural language tasks using few-shot learning, which drastically reduces the number of task-specific training examples needed to adapt the model to a particular application. To further our understanding of the impact of scale on few-shot learning, we trained a 540-billion parameter, densely activated, Transformer language model, which we call Pathways Language Model PaLM. We trained PaLM on 6144 TPU v4 chips using Pathways, a new ML system which enables highly efficient training across multiple TPU Pods. We demonstrate continued benefits of scaling by achieving state-of-the-art few-shot learning results on hundreds of language understanding and generation benchmarks. On a number of these tasks, PaLM 540B achieves breakthrough performance, outperforming the finetuned state-of-the-art on a suite of multi-step reasoning tasks, and outperforming average human performance on the recently released BIG-bench benchmark. A significant number of BIG-bench tasks showed discontinuous improvements from model scale, meaning that performance steeply increased as we scaled to our largest model. PaLM also has strong capabilities in multilingual tasks and source code generation, which we demonstrate on a wide array of benchmarks. We additionally provide a comprehensive analysis on bias and toxicity, and study the extent of training data memorization with respect to model scale. Finally, we discuss the ethical considerations related to large language models and discuss potential mitigation strategies.
hub tools
citation-role summary
background 54
method 2
other 2
baseline 1
citation-polarity summary
claims ledger
- abstract Large language models have been shown to achieve remarkable performance across a variety of natural language tasks using few-shot learning, which drastically reduces the number of task-specific training examples needed to adapt the model to a particular application. To further our understanding of the impact of scale on few-shot learning, we trained a 540-billion parameter, densely activated, Transformer language model, which we call Pathways Language Model PaLM. We trained PaLM on 6144 TPU v4 chips using Pathways, a new ML system which enables highly efficient training across multiple TPU Pod
authors
co-cited works
representative citing papers
DSPy compiles short declarative programs into LM pipelines that self-optimize and outperform both standard few-shot prompting and expert-written chains on math, retrieval, and QA tasks.
AgentBench is a new multi-environment benchmark showing commercial LLMs outperform open-source models up to 70B parameters in agent tasks mainly due to better long-term reasoning and instruction following.
Tree of Thoughts enables language models to solve complex planning tasks by generating, evaluating, and searching over coherent intermediate thoughts in a tree, raising Game of 24 success from 4% to 74% with GPT-4.
Transformers performing in-context learning implicitly implement gradient descent, ridge regression, and least-squares predictors for linear models, with behavior shifting based on model depth, width, and data noise.
PAL improves few-shot reasoning accuracy by having LLMs generate executable programs rather than text-based chains of thought, outperforming much larger models on math and logic benchmarks.
Language models generate robot policy code from natural language commands via few-shot prompting, enabling spatial-geometric reasoning, generalization, and precise control on real robots.
GPT-3 can learn to express well-calibrated uncertainty about its answers using natural language phrases rather than logits.
EntropyScan detects backdoored LVLMs by quantifying structural anomalies in visual attention distributions on benign samples via Tsallis entropy and reference-anchored Z-score normalization.
LLM tasks are supported by multiple distinct circuits rather than unique mechanisms, demonstrated via Overlap-Aware Sheaf Repulsion and the Distributive Dense Circuit Hypothesis.
VORT assigns learnable fractional orders to tokens and approximates their power-law retention kernels via sum-of-exponentials for efficient long-range dependency modeling in transformers.
Parallel-SFT mixes parallel programs across languages during SFT to produce more transferable RL initializations, yielding better zero-shot generalization to unseen programming languages.
SMC forgets its initial condition geometrically in the jump chain and as 1/ℓ in continuous genetic distance, justifying independent-locus approximations.
ATLAS is the first silicon-validated simulation framework for 3D-DRAM LLM accelerators, achieving under 8.57% error and over 97% correlation with real hardware while supporting design exploration.
Users treat human delegation for long tasks as a flexible compass but AI delegation as rigid railway tracks due to perceived AI limitations in inference and judgment.
A recurrent-depth architecture enables language models to improve reasoning performance by iterating computation in latent space, achieving gains equivalent to much larger models on benchmarks.
KV cache compression causes task-dependent degradation in high-density reasoning due to disrupted CoT links; ShotKV mitigates this by preserving few-shot examples as indivisible semantic units through phase separation, delivering 9-18% accuracy gains and 11% latency reduction.
Moshi is the first real-time full-duplex spoken large language model that casts dialogue as speech-to-speech generation using parallel audio streams and an inner monologue of time-aligned text tokens.
Medusa augments LLMs with multiple decoding heads and tree-based attention to predict and verify several tokens in parallel, yielding 2.2-3.6x inference speedup via two fine-tuning regimes.
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.
LRM is a large transformer that predicts a NeRF directly from a single image after training on a million-object multi-view dataset.
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.
Set-of-Mark prompting marks segmented image regions with alphanumerics and masks to let GPT-4V achieve state-of-the-art zero-shot results on referring expression comprehension and segmentation benchmarks like RefCOCOg.
UniSim learns a universal real-world simulator from orchestrated diverse datasets, enabling zero-shot deployment of policies trained purely in simulation.
citing papers explorer
-
MMMU: A Massive Multi-discipline Multimodal Understanding and Reasoning Benchmark for Expert AGI
MMMU provides 11.5K heterogeneous college-level multimodal questions that current models solve at 56-59% accuracy, establishing a new standard for expert multimodal evaluation.
-
DSPy: Compiling Declarative Language Model Calls into Self-Improving Pipelines
DSPy compiles short declarative programs into LM pipelines that self-optimize and outperform both standard few-shot prompting and expert-written chains on math, retrieval, and QA tasks.
-
AgentBench: Evaluating LLMs as Agents
AgentBench is a new multi-environment benchmark showing commercial LLMs outperform open-source models up to 70B parameters in agent tasks mainly due to better long-term reasoning and instruction following.
-
Tree of Thoughts: Deliberate Problem Solving with Large Language Models
Tree of Thoughts enables language models to solve complex planning tasks by generating, evaluating, and searching over coherent intermediate thoughts in a tree, raising Game of 24 success from 4% to 74% with GPT-4.
-
What learning algorithm is in-context learning? Investigations with linear models
Transformers performing in-context learning implicitly implement gradient descent, ridge regression, and least-squares predictors for linear models, with behavior shifting based on model depth, width, and data noise.
-
PAL: Program-aided Language Models
PAL improves few-shot reasoning accuracy by having LLMs generate executable programs rather than text-based chains of thought, outperforming much larger models on math and logic benchmarks.
-
Code as Policies: Language Model Programs for Embodied Control
Language models generate robot policy code from natural language commands via few-shot prompting, enabling spatial-geometric reasoning, generalization, and precise control on real robots.
-
Teaching Models to Express Their Uncertainty in Words
GPT-3 can learn to express well-calibrated uncertainty about its answers using natural language phrases rather than logits.
-
EntropyScan: Towards Model-level Backdoor Detection in LVLMs via Visual Attention Entropy
EntropyScan detects backdoored LVLMs by quantifying structural anomalies in visual attention distributions on benign samples via Tsallis entropy and reference-anchored Z-score normalization.
-
All Circuits Lead to Rome: Rethinking Functional Anisotropy in Circuit and Sheaf Discovery for LLMs
LLM tasks are supported by multiple distinct circuits rather than unique mechanisms, demonstrated via Overlap-Aware Sheaf Repulsion and the Distributive Dense Circuit Hypothesis.
-
VORT: Adaptive Power-Law Memory for NLP Transformers
VORT assigns learnable fractional orders to tokens and approximates their power-law retention kernels via sum-of-exponentials for efficient long-range dependency modeling in transformers.
-
Parallel-SFT: Improving Zero-Shot Cross-Programming-Language Transfer for Code RL
Parallel-SFT mixes parallel programs across languages during SFT to produce more transferable RL initializations, yielding better zero-shot generalization to unseen programming languages.
-
Rates of forgetting for the sequentially Markov coalescent
SMC forgets its initial condition geometrically in the jump chain and as 1/ℓ in continuous genetic distance, justifying independent-locus approximations.
-
A Full-Stack Performance Evaluation Infrastructure for 3D-DRAM-based LLM Accelerators
ATLAS is the first silicon-validated simulation framework for 3D-DRAM LLM accelerators, achieving under 8.57% error and over 97% correlation with real hardware while supporting design exploration.
-
Compass vs Railway Tracks: Unpacking User Mental Models for Communicating Long-Horizon Work to Humans vs. AI
Users treat human delegation for long tasks as a flexible compass but AI delegation as rigid railway tracks due to perceived AI limitations in inference and judgment.
-
Scaling up Test-Time Compute with Latent Reasoning: A Recurrent Depth Approach
A recurrent-depth architecture enables language models to improve reasoning performance by iterating computation in latent space, achieving gains equivalent to much larger models on benchmarks.
-
Semantic Integrity Matters: Benchmarking and Preserving High-Density Reasoning in KV Cache Compression
KV cache compression causes task-dependent degradation in high-density reasoning due to disrupted CoT links; ShotKV mitigates this by preserving few-shot examples as indivisible semantic units through phase separation, delivering 9-18% accuracy gains and 11% latency reduction.
-
Moshi: a speech-text foundation model for real-time dialogue
Moshi is the first real-time full-duplex spoken large language model that casts dialogue as speech-to-speech generation using parallel audio streams and an inner monologue of time-aligned text tokens.
-
Medusa: Simple LLM Inference Acceleration Framework with Multiple Decoding Heads
Medusa augments LLMs with multiple decoding heads and tree-based attention to predict and verify several tokens in parallel, yielding 2.2-3.6x inference speedup via two fine-tuning regimes.
-
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.
-
LRM: Large Reconstruction Model for Single Image to 3D
LRM is a large transformer that predicts a NeRF directly from a single image after training on a million-object multi-view dataset.
-
Detecting Pretraining Data from Large Language Models
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.
-
Set-of-Mark Prompting Unleashes Extraordinary Visual Grounding in GPT-4V
Set-of-Mark prompting marks segmented image regions with alphanumerics and masks to let GPT-4V achieve state-of-the-art zero-shot results on referring expression comprehension and segmentation benchmarks like RefCOCOg.
-
Learning Interactive Real-World Simulators
UniSim learns a universal real-world simulator from orchestrated diverse datasets, enabling zero-shot deployment of policies trained purely in simulation.
-
Language Model Beats Diffusion -- Tokenizer is Key to Visual Generation
A new shared video-image tokenizer enables large language models to surpass diffusion models on standard visual generation benchmarks.
-
C-Pack: Packed Resources For General Chinese Embeddings
C-Pack releases a new Chinese embedding benchmark, large training dataset, and optimized models that outperform priors by up to 10% on C-MTEB while also delivering English SOTA results.
-
Efficient Memory Management for Large Language Model Serving with PagedAttention
PagedAttention achieves near-zero waste in LLM key-value cache memory and enables 2-4x higher serving throughput than prior systems.
-
VoxPoser: Composable 3D Value Maps for Robotic Manipulation with Language Models
VoxPoser uses LLMs to compose 3D value maps via VLM interaction for model-based synthesis of robust robot trajectories on open-set language-specified manipulation tasks.
-
RepoBench: Benchmarking Repository-Level Code Auto-Completion Systems
RepoBench is a new benchmark with retrieval, completion, and pipeline tasks to evaluate code auto-completion systems on entire repositories instead of single files.
-
Direct Preference Optimization: Your Language Model is Secretly a Reward Model
DPO derives the optimal policy directly from human preferences via a reparameterized reward model, solving the RLHF objective with only a binary classification loss and no sampling or separate reward model.
-
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
-
RWKV: Reinventing RNNs for the Transformer Era
RWKV uses a linear attention mechanism to deliver Transformer-level performance with RNN-style inference efficiency, demonstrated at up to 14 billion parameters.
-
LIMA: Less Is More for Alignment
Fine-tuning a 65B model on 1,000 high-quality examples produces output that humans rate as good as or better than GPT-4 in 43% of cases, indicating most capabilities come from pretraining.
-
LLM+P: Empowering Large Language Models with Optimal Planning Proficiency
LLM+P lets LLMs solve planning problems optimally by converting them to PDDL for classical planners and back to natural language.
-
Visual Instruction Tuning
LLaVA is trained on GPT-4 generated visual instruction data to achieve 85.1% relative performance to GPT-4 on synthetic multimodal tasks and 92.53% accuracy on Science QA.
-
Segment Anything
A promptable model trained on 1B masks achieves competitive zero-shot segmentation performance across tasks and is released publicly with its dataset.
-
Language Is Not All You Need: Aligning Perception with Language Models
Kosmos-1 shows strong zero-shot and few-shot results on language tasks, image captioning, visual QA, OCR-free document understanding, and image recognition guided by text instructions.
-
Neural Codec Language Models are Zero-Shot Text to Speech Synthesizers
VALL-E is a neural codec language model trained on 60K hours of speech that performs zero-shot TTS, synthesizing natural speech that matches an unseen speaker's voice, emotion, and environment from a 3-second prompt.
-
Fast Inference from Transformers via Speculative Decoding
Speculative decoding accelerates exact sampling from large autoregressive models by 2-3x on T5-XXL by running smaller approximation models in parallel to propose token sequences that the large model then verifies in batches while preserving the original output distribution.
-
Program of Thoughts Prompting: Disentangling Computation from Reasoning for Numerical Reasoning Tasks
PoT prompting improves numerical reasoning by having language models write programs executed by a computer instead of performing calculations in natural language chains of thought, with an average 12% gain over CoT.
-
Mass-Editing Memory in a Transformer
MEMIT scales direct memory editing in transformers from single facts to thousands of associations by optimizing MLP weight updates.
-
PaLI: A Jointly-Scaled Multilingual Language-Image Model
PaLI jointly scales a 4B-parameter vision transformer with language models on a new 10B multilingual image-text dataset to reach state-of-the-art results on vision-language tasks while keeping a simple modular design.
-
A Generalist Agent
Gato is a multi-modal, multi-task, multi-embodiment generalist policy using one transformer network to handle text, vision, games, and robotics tasks.
-
OPT: Open Pre-trained Transformer Language Models
OPT releases open decoder-only transformers up to 175B parameters that match GPT-3 performance at one-seventh the carbon cost, along with code and training logs.
-
Flamingo: a Visual Language Model for Few-Shot Learning
Flamingo models reach new state-of-the-art few-shot results on image and video tasks by bridging frozen vision and language models with cross-attention layers trained on interleaved web-scale data.
-
Socratic Models: Composing Zero-Shot Multimodal Reasoning with Language
Socratic Models compose zero-shot multimodal reasoning by prompting pretrained language and vision models to exchange information and enable new capabilities without finetuning.
-
Omni-Perception Policy Optimization for Multimodal Emotion Reasoning
A cue-coverage reward plus a modality-token KL penalty makes multimodal emotion-reasoning models cite more real visual/audio evidence and hallucinate less, with reported SoTA on emotion benchmarks.
-
Variable-Width Transformers
×-shaped variable-width transformers outperform parameter-matched uniform baselines on language modeling loss with 22% fewer FLOPs and 15% smaller KV cache.
-
VTI-CoT: Visual-Textual Interleaved Chain of Thought for Video Reasoning
VTI-CoT proposes a visual-textual interleaved chain-of-thought method for video reasoning, built via automated annotation and OCR compression, claiming SOTA performance and better training efficiency on same-scale models.
-
Finite Certificates for In-Context Determinacy and a Threshold Theory of Emergence in Language Models
Proves row-space criterion for finite determinacy in linear finite-field tasks, NP-completeness of minimal forcing subcontext, and anti-mirage theorem separating threshold metrics from semantic confidence via Keisler measures.