Flat minima are illusory; generalization is driven by weakness, a reparameterization-invariant measure of compatible completions that predicts performance better than sharpness on MNIST and Fashion-MNIST.
super hub Canonical reference
Language Models are Few-Shot Learners
Canonical reference. 76% of citing Pith papers cite this work as background.
358
Pith papers citing it
Background
76% of classified citations
abstract
Recent work has demonstrated substantial gains on many NLP tasks and benchmarks by pre-training on a large corpus of text followed by fine-tuning on a specific task. While typically task-agnostic in architecture, this method still requires task-specific fine-tuning datasets of thousands or tens of thousands of examples. By contrast, humans can generally perform a new language task from only a few examples or from simple instructions - something which current NLP systems still largely struggle to do. Here we show that scaling up language models greatly improves task-agnostic, few-shot performance, sometimes even reaching competitiveness with prior state-of-the-art fine-tuning approaches. Specifically, we train GPT-3, an autoregressive language model with 175 billion parameters, 10x more than any previous non-sparse language model, and test its performance in the few-shot setting. For all tasks, GPT-3 is applied without any gradient updates or fine-tuning, with tasks and few-shot demonstrations specified purely via text interaction with the model. GPT-3 achieves strong performance on many NLP datasets, including translation, question-answering, and cloze tasks, as well as several tasks that require on-the-fly reasoning or domain adaptation, such as unscrambling words, using a novel word in a sentence, or performing 3-digit arithmetic. At the same time, we also identify some datasets where GPT-3's few-shot learning still struggles, as well as some datasets where GPT-3 faces methodological issues related to training on large web corpora. Finally, we find that GPT-3 can generate samples of news articles which human evaluators have difficulty distinguishing from articles written by humans. We discuss broader societal impacts of this finding and of GPT-3 in general.
hub tools
citation-role summary
background 70
method 9
dataset 2
baseline 1
citation-polarity summary
claims ledger
- abstract Recent work has demonstrated substantial gains on many NLP tasks and benchmarks by pre-training on a large corpus of text followed by fine-tuning on a specific task. While typically task-agnostic in architecture, this method still requires task-specific fine-tuning datasets of thousands or tens of thousands of examples. By contrast, humans can generally perform a new language task from only a few examples or from simple instructions - something which current NLP systems still largely struggle to do. Here we show that scaling up language models greatly improves task-agnostic, few-shot performan
authors
co-cited works
representative citing papers
The SDE benchmark shows LLMs lag on scientific discovery tasks relative to general science tests, with diminishing scaling returns and shared weaknesses across models.
A Lean-verified multi-agent system produces a catalogue of 14,116 quantum codes with transversal diagonal gates for small parameters, extracts infinite families, and resolves specific distance-3 cases with constructions and no-go proofs.
LLaDA is a scalable diffusion-based language model that matches autoregressive LLMs like LLaMA3 8B on tasks and surpasses GPT-4o on reversal poem completion.
Generative agents with memory streams, reflection, and planning using LLMs exhibit believable individual and emergent social behaviors in a simulated town.
Task vectors from weight differences allow arithmetic operations to edit pre-trained models, improving multiple tasks simultaneously and enabling analogical inference on unseen tasks.
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.
Training language models to generate intermediate computation steps on a scratchpad enables them to perform multi-step tasks such as long addition and arbitrary program execution that they otherwise fail at.
Decision Transformer casts RL as autoregressive sequence modeling conditioned on desired returns, past states and actions, matching or exceeding offline RL baselines on Atari, Gym and Key-to-Door tasks.
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.
GPT-f, a transformer-based prover for Metamath, generated new short proofs that were accepted into the main library—the first such contribution from a deep-learning system.
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.
APEX4 co-designs pure INT4 GEMM kernels with ρ-aware granularity adaptation to deliver up to 2.09× end-to-end speedup on GPUs with low ρ while keeping LLaMA-2-70B perplexity within 0.63 of FP16.
Face-Feature Tuning is a label-free logit remapping method that reduces FPR/TPR gaps across groups in deepfake detection while preserving overall accuracy.
The authors introduce a three-level formality spectrum (informal, casual, formal) and the 3LF dataset to correct supervision misalignment in formality transfer, reporting large gains in informal-to-formal performance on models including GPT variants.
UA-Legal-Bench is a new five-task benchmark for Ukrainian legal reasoning that demonstrates task-dependent few-shot prompting effects and the need for macro-F1 over accuracy on imbalanced classes.
LOSCAR-SGD combines local updates, sparse model averaging, and communication-computation overlap with a delay-corrected merge rule, providing convergence rates for smooth non-convex objectives under worker heterogeneity.
Ringmaster LMO extends delay-thresholding from ASGD to LMO-based momentum updates, providing convergence guarantees under (L0, L1)-smoothness and time-complexity bounds that recover optimal rates in the Euclidean case.
Formalizes Reasoning Portability (RP) and proposes RDB-CL to modulate per-sample KL regularization in RLVR for MLLM continual learning, achieving +12.0% Last accuracy over vanilla RLVR baseline by preserving reusable reasoning on high-RP samples.
LLM tasks are supported by multiple distinct circuits rather than unique mechanisms, demonstrated via Overlap-Aware Sheaf Repulsion and the Distributive Dense Circuit Hypothesis.
Fast-Slow Training uses context optimization as fast weights alongside parameter updates as slow weights to achieve up to 3x better sample efficiency, higher performance, and less catastrophic forgetting than standard RL in continual LLM learning.
The global empirical NTK for finite-width networks has a universal Kronecker-core form that makes it structurally low-rank and biases gradient descent toward dominant modes of joint input-hidden activity.
Inkheart SGD and M4 use bidirectional compression to achieve time complexities in distributed SGD that improve with worker count n and surpass prior lower bounds under a necessary structural assumption.
Topology-enhanced alignment via persistent homology on trajectories outperforms standard SFT and DPO baselines on preference metrics for LLMs.
citing papers explorer
-
Scalable Distributed Stochastic Optimization via Bidirectional Compression: Beyond Pessimistic Limits
Inkheart SGD and M4 use bidirectional compression to achieve time complexities in distributed SGD that improve with worker count n and surpass prior lower bounds under a necessary structural assumption.
-
From Static Analysis to Audience Dissemination: A Training-Free Multimodal Controversy Detection Multi-Agent Framework
AuDisAgent reformulates multimodal controversy detection as a dynamic audience dissemination process using screening, panel discussion, and arbitration agents, plus comment bootstrapping, and reports outperforming prior static methods on a public dataset.
-
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.
-
Hierarchical Text-Conditional Image Generation with CLIP Latents
A hierarchical prior-decoder model using CLIP latents generates more diverse text-conditional images than direct methods while preserving photorealism and caption fidelity.
-
Prefix-Tuning: Optimizing Continuous Prompts for Generation
Prefix-tuning matches or exceeds fine-tuning on NLG tasks by optimizing a continuous prefix using 0.1% of parameters while keeping the LM frozen.
-
GEM: Graph-Enhanced Mixture-of-Experts with ReAct Agents for Dialogue State Tracking
GEM achieves 65.19% joint goal accuracy on MultiWOZ 2.2 by routing between a graph neural network expert for dialogue structure and a T5 expert for sequences, plus ReAct agents for value generation, outperforming prior SOTA methods.
-
Evaluating Multi-Hop Reasoning in RAG Systems: A Comparison of LLM-Based Retriever Evaluation Strategies
CARE, a context-aware LLM judge, outperforms standard methods when evaluating multi-hop retrieval quality in RAG systems.
-
Representation Engineering: A Top-Down Approach to AI Transparency
Representation engineering uses population-level representations in deep neural networks to monitor and manipulate cognitive phenomena like honesty and harmlessness, providing simple effective baselines for LLM safety.
-
IUQ: Interrogative Uncertainty Quantification for Long-Form Large Language Model Generation
IUQ quantifies claim-level uncertainty in long-form LLM generation by combining inter-sample consistency and intra-sample faithfulness through an interrogate-then-respond approach and outperforms baselines on two datasets.