Pythia releases 16 identically trained LLMs with full checkpoints and data tools to study training dynamics, scaling, memorization, and bias in language models.
super hub Canonical reference
Emergent Abilities of Large Language Models
Canonical reference. 86% of citing Pith papers cite this work as background.
128
Pith papers citing it
Background
86% of classified citations
abstract
Scaling up language models has been shown to predictably improve performance and sample efficiency on a wide range of downstream tasks. This paper instead discusses an unpredictable phenomenon that we refer to as emergent abilities of large language models. We consider an ability to be emergent if it is not present in smaller models but is present in larger models. Thus, emergent abilities cannot be predicted simply by extrapolating the performance of smaller models. The existence of such emergence implies that additional scaling could further expand the range of capabilities of language models.
hub tools
citation-role summary
background 33
baseline 2
citation-polarity summary
claims ledger
- abstract Scaling up language models has been shown to predictably improve performance and sample efficiency on a wide range of downstream tasks. This paper instead discusses an unpredictable phenomenon that we refer to as emergent abilities of large language models. We consider an ability to be emergent if it is not present in smaller models but is present in larger models. Thus, emergent abilities cannot be predicted simply by extrapolating the performance of smaller models. The existence of such emergence implies that additional scaling could further expand the range of capabilities of language model
authors
co-cited works
representative citing papers
Grokking arises from gradual amplification of a Fourier-based circuit in the weights followed by removal of memorizing components.
GPT-2 small solves indirect object identification via a circuit of 26 attention heads organized into seven functional classes discovered through causal interventions.
Token loss trajectories follow localized sigmoids whose learning-time spectrum quantitatively reconstructs scaling-law derivatives on T, D, and M axes and enables faster training via distribution reshaping.
Self-evolving rubric with anti-gaming fitness reveals that objective capability scaling fails to transfer to subjective LLM behaviors, with advice-restraint as the universal lowest dimension that can regress.
A multi-agent pipeline iteratively refines topology optimization outputs to match natural language preferences for branched structures, achieving 60% success rate across replicates in cantilever and phone-stand tasks.
LLMs copy biased analyst ratings in investment decisions but a new detection method encourages independent reasoning and can improve stock return predictions beyond human levels.
Graphlets mined as structural tokens improve zero-shot inductive and transductive link prediction in knowledge graph foundation models across 51 diverse graphs.
A new 7x4 taxonomy organizes agentic AI security threats by architectural layer and persistence timescale, revealing under-explored upper layers and missing defenses after surveying 116 papers.
A 2D neural cellular automaton spontaneously self-organizes into a Proto-CKY representation that exhibits syntactic processing capabilities for context-free grammars when trained on membership problems.
PERCEIVE is the first bilingual benchmark integrating author content, reader emotions from comments, communication behavior, user attributes, and social graphs for personalized social media emotion understanding.
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.
LLM adoption in science follows a compressing inverted-U trajectory where release year predicts time-to-peak and lifespan better than model attributes.
Social dynamics in LLM collectives cause representative agents to make less accurate decisions as peer pressure increases through larger adversarial groups, more capable peers, longer arguments, and persuasive styles.
BoostTaxo introduces a boosting-style LLM framework for zero-shot taxonomy induction that uses hybrid candidate selection and constraint-aware calibration to achieve superior or comparable performance to prior methods on WordNet, DBLP, and SemEval-Sci benchmarks.
StackRepoQA shows LLMs reach only moderate accuracy on multi-file Java QA tasks, with gains from graph-based retrieval but frequent reliance on verbatim answer reproduction.
FactorEngine mines alpha factors as Turing-complete code via LLM-guided directional search, parameter separation, and a multi-agent pipeline that converts financial reports into executable programs, delivering higher IC/ICIR and Sharpe ratios than baselines in backtests.
Retrieval-augmented LLMs produce more cautious and guideline-aligned recommendations on cannabidiol for older adults than standalone models, demonstrated via automated evaluation on 64 diverse scenarios.
The ghost mechanism derives a 1D canonical model of abrupt learning in RNNs from ghost points of saddle-node bifurcations, predicting an inverse-power-law critical learning rate and gradient-based failure modes.
PolyMATH is a new 5,000-image benchmark where top MLLMs reach at most 41 percent accuracy on multi-modal mathematical reasoning, with ablation showing minimal gain from text over images.
Scaled vanilla autoregressive models based on Llama achieve 2.18 FID on ImageNet 256x256 image generation, beating popular diffusion models without visual inductive biases.
Introduces YesBut benchmark showing state-of-the-art multimodal models lag humans on interpreting humorous contradictions in comics.
DeepSeek-V2 delivers top-tier open-source LLM performance using only 21B active parameters by compressing the KV cache 93.3% and cutting training costs 42.5% via MLA and DeepSeekMoE.
GPT-4 LLM agents autonomously exploit 87% of tested one-day vulnerabilities when given CVE descriptions, far outperforming other models and tools.
citing papers explorer
-
Identification of quantum generative circuits with parallel quantum neural network
ParaQuanNet distinguishes eight quantum generative circuits via 99.5% accurate classification of their output data using parallel quantum embeddings and mutually unbiased measurements.
-
Multi-Agent Home Energy Management Assistant
HEMA is a multi-agent LLM system with analysis, knowledge, and control agents plus a self-consistency router that enables conversational home energy tasks, evaluated via LLM-simulated users on 23 metrics.
-
"The Whole Is Greater Than the Sum of Its Parts": A Compatibility-Aware Multi-Teacher CoT Distillation Framework
COMPACT adaptively fuses multi-teacher CoT supervisions using graph-based consensus, mutual-information adaptability, and loss-based difficulty metrics to improve small language model reasoning performance while mitigating catastrophic forgetting.
-
Large Language Model Agent for User-friendly Chemical Process Simulations
An LLM agent integrated with AVEVA Process Simulation via MCP enables natural language driven flowsheet analysis, optimization, and construction for chemical separation processes.
-
SONIC: Supersizing Motion Tracking for Natural Humanoid Whole-Body Control
Scaling motion tracking models along size, data volume, and compute produces a foundation model for natural, robust humanoid whole-body control with downstream uses in kinematic planning and vision-language-action models.
-
Scaling Laws Meet Model Architecture: Toward Inference-Efficient LLMs
A conditional scaling law fitted on over 200 models from 80M to 3B parameters identifies architectures that deliver up to 2.1% higher accuracy and 42% higher inference throughput than LLaMA-3.2 under the same training budget.
-
Video models are zero-shot learners and reasoners
Generative video models exhibit emergent zero-shot capabilities across perception, manipulation, and basic reasoning tasks.
-
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.
-
Multi-SpatialMLLM: Multi-Frame Spatial Understanding with Multi-Modal Large Language Models
Multi-SpatialMLLM integrates depth perception, visual correspondence, and dynamic perception into MLLMs via a 27M-sample MultiSPA dataset and benchmark, yielding gains on multi-frame spatial tasks.
-
Towards an AI co-scientist
A multi-agent AI system generates novel biomedical hypotheses that show promising experimental validation in drug repurposing for leukemia, new targets for liver fibrosis, and a bacterial gene transfer mechanism.
-
$\pi_0$: A Vision-Language-Action Flow Model for General Robot Control
π₀ is a vision-language-action flow model trained on diverse multi-platform robot data that supports zero-shot task performance, language instruction following, and efficient fine-tuning for dexterous tasks.
-
Jailbroken: How Does LLM Safety Training Fail?
LLM safety training fails due to competing objectives and mismatched generalization, enabling new jailbreaks that succeed on all unsafe prompts from red-teaming sets in GPT-4 and Claude.
-
H$_2$O: Heavy-Hitter Oracle for Efficient Generative Inference of Large Language Models
H2O evicts non-heavy-hitter tokens from the KV cache using a dynamic submodular policy, retaining recent and frequent-co-occurrence tokens to reduce memory while preserving accuracy.
-
Towards Expert-Level Medical Question Answering with Large Language Models
Med-PaLM 2 achieves 86.5% accuracy on MedQA and approaches or exceeds prior state-of-the-art on other medical QA benchmarks while receiving higher physician preference ratings than human answers on consumer questions.
-
HuggingGPT: Solving AI Tasks with ChatGPT and its Friends in Hugging Face
HuggingGPT is an agent system where ChatGPT plans and orchestrates calls to Hugging Face models to solve complex multi-modal AI tasks.
-
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.
-
ART: Automatic multi-step reasoning and tool-use for large language models
ART automatically generates multi-step reasoning programs with tool integration for LLMs, yielding substantial gains over few-shot and auto-CoT prompting on BigBench and MMLU while matching hand-crafted CoT on most tasks.
-
Large Language Models Are Human-Level Prompt Engineers
APE generates instruction candidates via LLM and selects the best by zero-shot performance of a second LLM, matching or beating human prompts on 19 of 24 NLP tasks.
-
Large Language Models Can Self-Improve
A 540B-parameter LLM improves reasoning performance on GSM8K, DROP, OpenBookQA, and ANLI-A3 by fine-tuning on self-generated high-confidence CoT solutions from unlabeled data.
-
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.
-
A Single Rewrite Suffices: Empirical Lessons from Production Skill Description Optimization
A single LLM rewrite of skill descriptions using false positive and negative cases matches manual optimization performance in production, with most other pipeline components adding little value.
-
Characterize Then Distill: Mechanistic Reasoning in Large Output Spaces
Reasoning in large output spaces proceeds via shortlisting then fine-grained reasoning; this characterization enables a mechanistic distillation strategy that outperforms standard distillation.
-
In-Context Reward Adaptation for Robust Preference Modeling
Transformer model with response-time auxiliary input adapts reward models to unseen human preference domains via in-context learning from demonstrations.
-
Beyond Scaling: Agents Are Heading to the Edge
Personal agents require edge deployment to preserve high-fidelity local context and zero-latency loops, as claimed through three structural shifts away from cloud-centric designs.
-
Unveiling Memorization-Generalization Coexistence: A Case Study on Arithmetic Tasks with Label Noise
Experiments on modular arithmetic with heavy label noise show that over-parameterized networks form a distributed internal generalization structure that can be extracted via frequency methods to achieve high accuracy despite 80% noise.
-
Agentic AIs Are the Missing Paradigm for Out-of-Distribution Generalization in Foundation Models
Agentic AI systems are required to overcome the parameter coverage ceiling that prevents foundation models from handling certain out-of-distribution cases.
-
Novelty-based Tree-of-Thought Search for LLM Reasoning and Planning
Novelty estimation via LLM prompts enables pruning in Tree-of-Thought search, reducing overall token usage on language planning benchmarks.
-
Optimized Deferral for Imbalanced Settings
MILD reformulates two-stage learning to defer as cost-sensitive learning over the input-expert domain and derives new margin-based losses with guarantees, yielding better performance than baselines on image classification and LLM routing tasks.
-
STELLAR-E: a Synthetic, Tailored, End-to-end LLM Application Rigorous Evaluator
STELLAR-E modifies the TGRT Self-Instruct framework to produce tailored synthetic LLM evaluation datasets that score an average 5.7% higher on LLM-as-a-judge metrics than existing language-specific benchmarks.
-
Evaluating LLM-Based Goal Extraction in Requirements Engineering: Prompting Strategies and Their Limitations
LLM pipeline with generation-critic feedback reaches 61% accuracy on low-level goal extraction from requirements documents and outperforms standalone few-shot prompting, yet remains best suited as an accelerator for manual work.
-
Cooperative Profiles Predict Multi-Agent LLM Team Performance in AI for Science Workflows
Cooperative profiles from behavioral economics games predict LLM team performance in AI-for-science workflows.
-
Absorber LLM: Harnessing Causal Synchronization for Test-Time Training
Absorber LLM introduces causal synchronization to absorb context into parameters for memory-efficient long-context LLM inference while preserving causal effects.
-
ARMove: Learning to Predict Human Mobility through Agentic Reasoning
ARMove is a transferable framework for human mobility prediction that combines agentic LLM reasoning, feature management, and large-small model synergy to outperform baselines on several metrics while improving interpretability and robustness.
-
LACE: Lattice Attention for Cross-thread Exploration
LACE enables concurrent reasoning paths in LLMs to interact via lattice attention and a synthetic training pipeline, raising accuracy more than 7 points over independent parallel search.
-
The Cartesian Cut in Agentic AI
LLM agents use a Cartesian split between learned prediction and engineered control, enabling modularity but creating sensitivity and bottlenecks unlike integrated biological systems.
-
Limits of Difficulty Scaling: Hard Samples Yield Diminishing Returns in GRPO-Tuned SLMs
GRPO tuning on SLMs shows diminishing returns from hard math samples, with easier subsets matching full performance using 45% fewer steps and GSM8K training outperforming MATH training on numeric subsets.
-
Elder-Sim: A Psychometrically Validated Platform for Personality-Stable Elderly Digital Twins
ELDER-SIM builds personality-stable elderly digital twins via LLM orchestration with OCEAN traits, Beck CBT diagrams, long-term memory, and LoRA fine-tuning on CHARLS data, validated by Cronbach's alpha 0.70-0.94 and ICC 0.85-0.96.
-
Combating the Memory Walls: Optimization Pathways for Long-Context Agentic LLM Inference
PLENA introduces a co-designed system with three optimization pathways for long-context agentic LLM inference, claiming up to 2.23x throughput over A100 and 4.04x energy efficiency.
-
Query Expansion in the Age of Pre-trained and Large Language Models: A Comprehensive Survey
A comprehensive survey that organizes query expansion methods in the PLM/LLM era along four design dimensions, synthesizes application patterns, and outlines future directions.
-
Improving LLM First-Token Predictions in Multiple-Choice Question Answering via Output Prefilling
Output prefilling with a structured prefix steers LLMs to produce cleaner first tokens in MCQA, raising accuracy and calibration over standard first-token probability.
-
Emerging Properties in Unified Multimodal Pretraining
BAGEL is a unified decoder-only model that develops emerging complex multimodal reasoning abilities after pretraining on large-scale interleaved data and outperforms prior open-source unified models.
-
TokUR: Token-Level Uncertainty Estimation for Large Language Model Reasoning
TokUR estimates token-level uncertainty via low-rank weight perturbations in LLMs, aggregates signals to correlate with correctness, and uses them to improve reasoning performance on math tasks.
-
Exploring Cross-lingual Latent Transplantation: Mutual Opportunities and Open Challenges
XTransplant empirically shows that cross-lingual latent transplantation yields mutual benefits for multilingual capability and cultural adaptability in LLMs, especially low-resource ones, while revealing underutilized model potential.
-
CHESS: Contextual Harnessing for Efficient SQL Synthesis
CHESS deploys four LLM agents to retrieve information, prune schemas, generate refined SQL candidates, and validate via unit tests, reporting up to 71.10% accuracy on BIRD with 83% fewer calls than leading proprietary baselines.
-
TrustLLM: Trustworthiness in Large Language Models
TrustLLM defines eight trustworthiness principles, creates a six-dimension benchmark, and evaluates 16 LLMs showing proprietary models generally lead but some open-source ones are close while over-calibration can hurt utility.
-
MobileVLM : A Fast, Strong and Open Vision Language Assistant for Mobile Devices
MobileVLM achieves on-par performance with much larger vision-language models on standard benchmarks while delivering state-of-the-art inference speeds of 21.5 tokens per second on Snapdragon 888 CPU and 65.3 on Jetson Orin GPU.
-
Trustworthy LLMs: a Survey and Guideline for Evaluating Large Language Models' Alignment
Survey organizes LLM trustworthiness into seven categories and 29 sub-categories, measures eight sub-categories on popular models, and finds that more aligned models generally score higher but with varying effectiveness.
-
StarCoder: may the source be with you!
StarCoderBase matches or beats OpenAI's code-cushman-001 on multi-language code benchmarks; the Python-fine-tuned StarCoder reaches 40% pass@1 on HumanEval while retaining other-language performance.
-
Reward-Free Code Alignment from Pretrained or Fine-Tuned LLM: Unpacking the Trade-offs for Code Generation
Empirical study on five LLMs finds pretrained-to-aligned paths yield bigger gains over baseline than finetuned-to-aligned paths, though absolute accuracy remains lower for pretrained starts.
-
Mechanistic Personality Analysis of LLMs Steering Personality via Latent Feature Interventions
Applies sparse autoencoders to locate and steer latent features for OCEAN personality traits in LLMs while preserving benchmark performance.