LUMOS: Tracing Parametric Knowledge from Training Data to Behavioral Outputs in LLMs
cs.AI
Submitted: 2026-10-02
Updated: 2026-10-12
Code: https://github.com/agilumos/lumos
Sources
- The Reversal Curse: LLMs trained on "A is B" fail to learn "B is A"
- Studying Large Language Model Generalization with Influence Functions
- Measuring Massive Multitask Language Understanding
- Discovering Latent Knowledge in Language Models Without Supervision
- BERT Rediscovers the Classical NLP Pipeline
- Understanding intermediate layers using linear classifier probes
- Causal Mediation Analysis for Interpreting Neural NLP: The Case of Gender Bias
- A Careful Examination of Large Language Model Performance on Grade School Arithmetic
- Linearity of Relation Decoding in Transformer Language Models
- Mind the Generation Process: Fine-Grained Confidence Estimation During LLM Generation
- UniKnow: A Unified Framework for Reliable Language Model Behavior across Parametric and External Knowledge
- Detecting Hallucinations in Large Language Model Generation: A Token Probability Approach
- Teaching Models to Express Their Uncertainty in Words
- DataComp-LM: In search of the next generation of training sets for language models
- OpenAlex: A fully-open index of scholarly works, authors, venues, institutions, and concepts
- ComparisonQA: Evaluating Factuality Robustness of LLMs Through Knowledge Frequency Control and Uncertainty
- CRaFT: An Explanation-Based Framework for Evaluating Cultural Reasoning in Multilingual Language Models
- Training Verifiers to Solve Math Word Problems
- GSM-Symbolic: Understanding the Limitations of Mathematical Reasoning in Large Language Models
- MARS: Meaning-Aware Response Scoring for Uncertainty Estimation in Generative LLMs
Related papers
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Agents' Last Exam
- The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
- MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
- Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems