AAAI 2025 accepted papers on Large language models

501 of the 3,486 papers accepted at AAAI 2025 (14.4%) are on Large language models, compared with 8.3% at AAAI 2024. They fall into 7 topics.

See AAAI 2025 in the interactive map

Working on something in this area? Describe your idea in scime atlas to see the closest accepted papers by meaning and how crowded the spot has become.

Topics and papers

Language, safety and interpretability · 154 papers

LLMEmb: Large Language Model Can Be a Good Embedding Generator for Sequential Recommendation (31 citations)
POI-Enhancer: An LLM-based Semantic Enhancement Framework for POI Representation Learning (22 citations)
Language Ranker: A Metric for Quantifying LLM Performance Across High and Low-Resource Languages (21 citations)
Codec Does Matter: Exploring the Semantic Shortcoming of Codec for Audio Language Model (20 citations)
Causal Prompting: Debiasing Large Language Model Prompting Based on Front-Door Adjustment (18 citations)
Bias Unveiled: Investigating Social Bias in LLM-Generated Code (17 citations)
MSE-Adapter: A Lightweight Plugin Endowing LLMs with the Capability to Perform Multimodal Sentiment Analysis and Emotion Recognition (17 citations)
Understanding Emotional Body Expressions via Large Language Models (16 citations)
DABL: Detecting Semantic Anomalies in Business Processes Using Large Language Models (13 citations)
Is LLMs Hallucination Usable? LLM-based Negative Reasoning for Fake News Detection (13 citations)
LLM-RG4: Flexible and Factual Radiology Report Generation Across Diverse Input Contexts (13 citations)
Multi-Turn Jailbreaking Large Language Models via Attention Shifting (13 citations)

and 142 more in the interactive map.

AI and society · 108 papers

and 96 more in the interactive map.

Retrieval and knowledge · 84 papers

TableBench: A Comprehensive and Complex Benchmark for Table Question Answering (28 citations)
Harnessing Large Language Models for Knowledge Graph Question Answering via Adaptive Multi-Aspect Retrieval-Augmentation (27 citations)
Debate on Graph: A Flexible and Reliable Reasoning Framework for Large Language Models (24 citations)
Citations and Trust in LLM Generated Responses (22 citations)
LightPROF: A Lightweight Reasoning Framework for Large Language Model on Knowledge Graph (22 citations)
RATT: A Thought Structure for Coherent and Correct LLM Reasoning (22 citations)
Improving Retrieval Augmented Language Model with Self-Reasoning (14 citations)
MaFeRw: Query Rewriting with Multi-Aspect Feedbacks for Retrieval-Augmented Large Language Models (13 citations)
Attributive Reasoning for Hallucination Diagnosis of Large Language Models (12 citations)
Improving Complex Reasoning over Knowledge Graph with Logic-Aware Curriculum Tuning (12 citations)
LEARN: Knowledge Adaptation from Large Language Model to Recommendation for Practical Industrial Application (12 citations)
BLADE: Enhancing Black-Box Large Language Models with Small Domain-Specific Models (11 citations)

and 72 more in the interactive map.

Agents, code and math · 77 papers

AnalogCoder: Analog Circuit Design via Training-Free Code Generation (55 citations)
Multi-Objective Evolution of Heuristic Using Large Language Model (26 citations)
LLM-Powered User Simulator for Recommender System (24 citations)
Agent4Edu: Generating Learner Response Data by Generative Agents for Intelligent Education Systems (21 citations)
CVE-LLM: Ontology-Assisted Automatic Vulnerability Evaluation Using Large Language Models (20 citations)
HSEvo: Elevating Automatic Heuristic Design with Diversity-Driven Harmony Search and Genetic Algorithm Using LLMs (20 citations)
CodeHalu: Investigating Code Hallucinations in LLMs via Execution-based Verification (18 citations)
MAGIC: Generating Self-Correction Guideline for In-Context Text-to-SQL (17 citations)
Evaluating Mathematical Reasoning Beyond Accuracy (16 citations)
Evolutionary Large Language Model for Automated Feature Transformation (14 citations)
Argumentative Large Language Models for Explainable and Contestable Claim Verification (13 citations)
Can LLMs Obfuscate Code? A Systematic Analysis of Large Language Models into Assembly Code Obfuscation (13 citations)

and 65 more in the interactive map.

Instruction and fine-tuning · 38 papers

SWIFT: A Scalable Lightweight Infrastructure for Fine-Tuning (40 citations)
Calibrating Large Language Models with Sample Consistency (19 citations)
XCOT: Cross-lingual Instruction Tuning for Cross-lingual Chain-of-Thought Reasoning (12 citations)
Bridging the Gap: Enhancing LLM Performance for Low-Resource African Languages with New Benchmarks, Fine-Tuning, and Cultural Adjustments (9 citations)
Scaling Trends for Data Poisoning in LLMs (7 citations)
Improving Factuality in Large Language Models via Decoding-Time Hallucinatory and Truthful Comparators (6 citations)
Unveiling the Impact of Coding Data Instruction Fine-Tuning on Large Language Models Reasoning (6 citations)
Simulate and Eliminate: Revoke Backdoors for Generative Large Language Models (5 citations)
Capability Instruction Tuning (4 citations)
Chain-of-Instructions: Compositional Instruction Tuning on Large Language Models (4 citations)
From Words to Worth: Newborn Article Impact Prediction with LLM (4 citations)
Gradient Weight-normalized Low-rank Projection for Efficient LLM Training (4 citations)

and 26 more in the interactive map.

Alignment and preferences · 24 papers

Evaluating the Evaluator: Measuring LLMs’ Adherence to Task Evaluation Instructions (16 citations)
Approximated Variational Bayesian Inverse Reinforcement Learning for Large Language Model Alignment (3 citations)
Controllable Protein Sequence Generation with LLM Preference Optimization (3 citations)
Maximizing Signal in Human-Model Preference Alignment (3 citations)
Robust Multi-Objective Preference Alignment with Online DPO (2 citations)
Align-Pro: A Principled Approach to Prompt Optimization for LLM Alignment (1 citation)
Aligning Large Language Models for Faithful Integrity Against Opposing Argument (1 citation)
Correcting Large Language Model Behavior via Influence Function (1 citation)
LEGEND: Leveraging Representation Engineering to Annotate Safety Margin for Preference Datasets (1 citation)
MAPLE: A Framework for Active Preference Learning Guided by Large Language Models (1 citation)
Multi-Reference Preference Optimization for Large Language Models (1 citation)
PRIORITY2REWARD: Incorporating Healthworker Preferences for Resource Allocation Planning (1 citation)

and 12 more in the interactive map.

Reasoning and chain of thought · 16 papers

and 4 more in the interactive map.