ICLR 2025PastLarge language models
ICLR 2025 Workshop on Foundation Models in the Wild
ICLR 2025 FM-Wild Workshop
- Submission deadline
- Feb 11, 2025, 11:59 UTCimported from OpenReview — check the website for extensions
- Submission portal
- OpenReview
- Notes
- Topics were auto-suggested and may be imprecise — edits welcome.
Accepted papers (102)
Fetched from OpenReview (v2) on 2026-06-10.
"Short-length" Adversarial Training Helps LLMs Defend "Long-length" Jailbreak Attacks: Theoretical and Empirical Evidence
Accelerating Unbiased LLM Evaluation via Synthetic Feedback
ACTIVATION STEERING IN NEURAL THEOREM PROVERS
Adjustment for Confounding using Pre-Trained Representations
AdvWave: Stealthy Adversarial Jailbreak Attack against Large Audio-Language Models
Agentic Multimodal AI for Hyper-Personalized B2B and B2C Advertising in Competitive Markets: An AI-Driven Competitive Advertising Framework
AgentTaxo: Dissecting and Benchmarking Token Distribution of LLM Multi-Agent Systems
All It Takes Is One Prompt: An Autonomous LLM-MA System
AppVLM: A Lightweight Vision Language Model for Online App Control
Are DeepSeek R1 And Other Reasoning Models More Faithful?
Aria-UI: Visual Grounding for GUI Instructions
Attacking Multimodal OS Agents with Malicious Image Patches
Automated Benchmark Generation for Repository-Level Coding Tasks
Automated Capability Discovery via Model Self-Exploration
AutoToM: Automated Bayesian Inverse Planning and Model Discovery for Open-ended Theory of Mind
Beautiful Images, Toxic Words: Understanding and Addressing Offensive Text in Generated Images
Beyond ID Bias: PCA-Guided Dropout for Robust Fine-tuning
Beyond Pixels: Enhancing LIME with Hierarchical Features and Segmentation Foundation Models
Bridging vision language model (VLM) evaluation gaps with a framework for scalable and cost-effective benchmark generation
Captured by Captions: On Memorization and its Mitigation in CLIP Models
CARROT: A Cost Aware Rate Optimal Router
Cheap and Effective Personalization of Foundation Language Models for Imitating a User's Writing Style
Co-optimizing Recommendation and Evaluation for LLM Selection
Cost-efficient Collaboration between On-device and Cloud Language Models
CROSS: Analyzing the Trade-offs in Long-Context Cross-lingual Retrieval
DASFormer: Self-supervised Pretraining for Earthquake Monitoring
DeltaProduct: Improving State-Tracking in Linear RNNs via Householder Products
Demystifying Long Chain-of-Thought Reasoning in LLMs
Detecting Covariate Shifts With Vision-Language Foundation Models
Diagnosing Robotics Systems Issues with Large Language Models -- A Case Study
Disentangling Sequence Memorization and General Capability in Large Language Models
Does Cross-Domain Pre-Training Truly Help Time-Series Foundation Models?
DP-GPL: DIFFERENTIALLY PRIVATE GRAPH PROMPT LEARNING
Efficient Backdoor Detection on Text-to-image Synthesis via Neuron Activation Variation
Emergent Misalignment: Narrow finetuning can produce broadly misaligned LLMs
Exploring LLM Agents for Cleaning Tabular Machine Learning Datasets
Faster, Cheaper, Better: Multi-Objective Hyperparameter Optimization for LLM and RAG Systems
Few-Shot Whole Slide Pathology Classification with Multi-Granular Vision-Language Models
FlipAttack: Jailbreak LLMs via Flipping
FocalLens: Instruction Tuning Enables Zero-Shot Conditional Image Representations
Focus on this, not that! Steering LLMs with Adaptive Feature Specification
Foundation Model-Based Data Selection for Dense Prediction Tasks
From Intuition to Understanding: Using AI Peers to Overcome Physics Misconceptions
G-Designer: Architecting Multi-agent Communication Topologies via Graph Neural Networks
Generalizing from SIMPLE to HARD Visual Reasoning: Can We Mitigate Modality Imbalance in VLMs?
Geneshift: Impact of different scenario shift on Jailbreaking LLM
GeoFT: Fine-tuning Foundation Models for Automated OSINT Geolocation
GuardReasoner: Towards Reasoning-based LLM Safeguards
Improving Your Model Ranking on Chatbot Arena by Vote Rigging
Inference Optimal VLMs Need Fewer Visual Tokens and More Parameters
Infinite Leagues Under the Sea: Realistic 3D Underwater Terrain Generation Augmented by Visual Foundation Models
KnowGuard: Robust Reasoning Enabled LLM Guardrail via Knowledge-Enhanced Logical Reasoning
KnowHalu: Hallucination Detection via Multi-Form Knowledge Based Factual Checking
Latent Representation Encoding and Multimodal Biomarkers for Post-Stroke Speech Assessment
Leveraging the true depth of LLMs
MASQUE: Diffusion-Based Localized Adversarial Makeup for Facial Privacy
Measuring In-Context Computation Complexity via Hidden State Prediction
MetaSC: Test-Time Safety Specification Optimization for Language Models
MITIGATING CACHE NOISE IN TEST-TIME ADAPTATION FOR LARGE VISION-LANGUAGE MODELS
MLLM CAN SEE? DYNAMIC CORRECTION DECODING FOR HALLUCINATION MITIGATION
MMInference: Accelerating Pre-filling for Long-Context Visual Language Models via Modality-Aware Permutation Sparse Attention
MoA: Mixture of Sparse Attention for Automatic Large Language Model Compression
Multi-Hypothesis Spatial Foundation Model: Rethinking and Decoupling Depth Ambiguity via Laplacian Visual Prompting
Narrowing Class-Wise Robustness Gaps in Adversarial Training
Navigating the Designs of Privacy-Preserving Fine-tuning for Large Language Models
OctoTools: An Agentic Framework with Extensible Tools for Complex Reasoning
OpenRAG: Optimizing RAG End-to-End via In-Context Retrieval Learning
Optimizing Test-Time Compute via Meta Reinforcement Finetuning
PARSE-Ego4D: Personal Action Recommendation Suggestions for Egocentric Videos
PhysBench: Benchmarking and Enhancing Vision-Language Models for Physical World Understanding
Policy-Agnostic RL: Offline RL and Online RL Fine-Tuning of Any Class and Backbone
Privacy Auditing for Large Language Models with Natural Identifiers
Reasoning Without Self-Doubt: More Efficient Chain-of-Thought Through Certainty Probing
ReFocus: Visual Editing as a Chain of Thought for Structured Image Understanding
Relevance Isn't All You Need: Scaling RAG Systems With Inference-Time Compute Via Multi-Criteria Reranking
Reliable and Efficient Amortized Model-based Evaluation
Risks and Safety Considerations for Foundation Model-based Autonomous Agents' Interaction with the Environment
RoboMorph: Evolving Robot Morphology using Large Language Models
SafeMERGE: Preserving Safety Alignment in Fine-Tuned Large Language Models via Selective Layer-Wise Model Merging
SafeWatch: An Efficient Safety-Policy Following Video Guardrail Model with Transparent Explanations
SAM2-Adapter: Evaluating & Adapting Segment Anything 2 in Downstream Tasks: Camouflage, Shadow, Medical Image Segmentation, and More
SGBD: Sharpness-Aware Mirror Gradient with BLIP-Based Denoising for Robust Multimodal Product Recommendation
Shh, don't say that! Domain Certification in LLMs
ShieldAgent: Shielding Agents via Verifiable Safety Policy Reasoning
Simple is Effective: The Roles of Graphs and Large Language Models in Knowledge-Graph-Based Retrieval-Augmented Generation
StochasTok: Improving Fine-Grained Subword Understanding in LLMs
Surgical-LVLM: Learning to Adapt Large Vision-Language Model for Grounded Visual Question Answering in Robotic Surgery
Tiled Flash Linear Attention: More Efficient Linear RNN and xLSTM Kernels
Toward Trustworthy Neural Program Synthesis
Towards Universal Offline Black-Box Optimization via Learning String Embedding Space
TPP-LLM: Modeling Temporal Point Processes by Efficiently Fine-Tuning Large Language Models
Tradeoffs Between Alignment and Helpfulness in Language Models with Steering Methods
Understanding (Un)Reliability of Steering Vectors in Language Models
Unisolver: PDE-Conditional Transformers Are Universal Neural PDE Solvers
Unlocking Post-hoc Dataset Inference with Synthetic Data
VisR-Bench: A Visual Retrieval Benchmark for Visually-Rich Documents
WABER: Evaluating Reliability and Efficiency of Web Agents with Existing Benchmarks
Why Foundation Models Struggle with Cross-Modal Context
Wider or Deeper? Scaling LLM Inference-Time Compute with Adaptive Branching Tree Search
Words or Vision: Do Vision-Language Models Have Blind Faith in Text?
WorkflowAgent: Towards Specialized Web Agents Using Production-Scale Workflow Data
xLSTM 7B: A Recurrent LLM for Fast and Efficient Inference