in-context learning
In-context learning allows models, particularly large language models, to adjust their responses based on specific prompts or contexts provided during inference, without requiring additional training or fine-tuning. It enables the model to exhibit flexible behavior in response to varying inputs.
- A Closer Look at TabPFN v2: Understanding Its Strengths and Extending Its Capabilities
- Any Large Language Model Can Be a Reliable Judge: Debiasing with a Reasoning-based Bias Detector
- Attractive Metadata Attack: Inducing LLM Agents to Invoke Malicious Tools
- Axial Neural Networks for Dimension-Free Foundation Models
- BAM-ICL: Causal Hijacking In-Context Learning with Budgeted Adversarial Manipulation
- Breaking the Gradient Barrier: Unveiling Large Language Models for Strategic Classification
- Bridging Sign and Spoken Languages: Pseudo Gloss Generation for Sign Language Translation
- CCL: Causal-aware In-context Learning for Out-of-Distribution Generalization
- CHASM: Unveiling Covert Advertisements on Chinese Social Media
- Causal Head Gating: A Framework for Interpreting Roles of Attention Heads in Transformers
- ConTextTab: A Semantics-Aware Tabular In-Context Learner
- Counterfactual reasoning: an analysis of in-context emergence
- Direct Numerical Layout Generation for 3D Indoor Scene Synthesis via Spatial Reasoning
- Disentangling Latent Shifts of In-Context Learning with Weak Supervision
- Do different prompting methods yield a common task representation in language models?
- Do-PFN: In-Context Learning for Causal Effect Estimation
- ENMA: Tokenwise Autoregression for Continuous Neural PDE Operators
- EquiTabPFN: A Target-Permutation Equivariant Prior Fitted Network
- Hierarchical Demonstration Order Optimization for Many-shot In-Context Learning
- How Data Mixing Shapes In-Context Learning: Asymptotic Equivalence for Transformers with MLPs
- In-Context Learning Strategies Emerge Rationally
- In-context Learning of Linear Dynamical Systems with Transformers: Approximation Bounds and Depth-separation
- LLM Meeting Decision Trees on Tabular Data
- Language Models Are Capable of Metacognitive Monitoring and Control of Their Internal Activations
- Large Language Diffusion Models
- Large Language Diffusion Models
- Large Language Models as Model Organisms for Human Associative Learning
- Learning World Models for Interactive Video Generation
- Learning to Learn with Contrastive Meta-Objective
- Learning to Learn with Contrastive Meta-Objective
- Learning to Rank for In-Context Example Retrieval
- Linear Transformers Implicitly Discover Unified Numerical Algorithms
- Memory Mosaics at scale
- Memory Mosaics at scale
- Meta-Learning an In-Context Transformer Model of Human Higher Visual Cortex
- Mitra: Mixed Synthetic Priors for Enhancing Tabular Foundation Models
- MotionRAG: Motion Retrieval-Augmented Image-to-Video Generation
- Nested Learning: The Illusion of Deep Learning Architectures
- OmniTalker: One-shot Real-time Text-Driven Talking Audio-Video Generation With Multimodal Style Mimicking
- On the Robustness of Transformers against Context Hijacking for Linear Classification
- Optimal Dynamic Regret by Transformers for Non-Stationary Reinforcement Learning
- Optimality and NP-Hardness of Transformers in Learning Markovian Dynamical Functions
- Optimization Inspired Few-Shot Adaptation for Large Language Models
- Pre-trained Large Language Models Learn to Predict Hidden Markov Models In-context
- Prior Forgetting and In-Context Overfitting
- ROVER: Recursive Reasoning Over Videos with Vision-Language Models for Embodied Tasks
- Reasoning Models Better Express Their Confidence
- RelationAdapter: Learning and Transferring Visual Relation with Diffusion Transformers
- Scaling Up Active Testing to Large Language Models
- Searching Latent Program Spaces
- Short-length Adversarial Training Helps LLMs Defend Long-length Jailbreak Attacks: Theoretical and Empirical Evidence
- TabDPT: Scaling Tabular Foundation Models on Real Data
- Technical Debt in In-Context Learning: Diminishing Efficiency in Long Context
- The Atlas of In-Context Learning: How Attention Heads Shape In-Context Retrieval Augmentation
- Theoretical Insights into In-context Learning with Unlabeled Data
- TiRex: Zero-Shot Forecasting Across Long and Short Horizons with Enhanced In-Context Learning
- Towards Predicting Any Human Trajectory In Context
- Trained Mamba Emulates Online Gradient Descent in In-Context Linear Regression
- Transformers are almost optimal metalearners for linear classification
- Unifying Attention Heads and Task Vectors via Hidden State Geometry in In-Context Learning
- Unlabeled Data Can Provably Enhance In-Context Learning of Transformers
- Unlocking SLM Potential for Data Analysis Code Generation via Non-Parametric Knowledge Distillation
- Variational Uncertainty Decomposition for In-Context Learning
- Vocabulary In-Context Learning in Transformers: Benefits of Positional Encoding
- What One Cannot, Two Can: Two-Layer Transformers Provably Represent Induction Heads on Any-Order Markov Chains
- When Thinking Fails: The Pitfalls of Reasoning for Instruction-Following in LLMs