7 papers
Belief-Calibrated Optimization: An Explicit World Model for Agentic Optimization
Yuhan Chen, Zhihua Tian, Mahavir Dabas +7
The performance of an LLM agent depends on the scaffold around a frozen model. A common way to improve that scaffold is to use a coding agent as an optimizer: it reads current scor…
Can Generalist Agents Automate Data Curation?
Feiyang Kang, Hanze Li, Adam Nguyen +5
Curating training data is among the most consequential yet labor-intensive parts of modern AI development: practitioners iteratively propose, implement, evaluate, and revise data p…
Memory-Induced Tool-Drift in LLM Agents
Mahavir Dabas, Jihyun Jeong, Ming Jin +1
LLM agents increasingly combine long-term memory for personalization with tool calling for taking actions. We study a failure of this combination: personality-driven preferences st…
Characterizing Model-Native Skills
Feiyang Kang, Mahavir Dabas, Myeongseob Ko +1
Skills are a natural unit for describing what a language model can do and how its behavior can be changed. However, existing characterizations rely on human-written taxonomies, tex…
Adversarial Déjà Vu: Jailbreak Dictionary Learning for Stronger Generalization to Unseen Attacks
Mahavir Dabas, Tran Huynh, Nikhil Reddy Billa +8
Large language models remain vulnerable to jailbreak attacks that bypass safety guardrails to elicit harmful outputs. Defending against novel jailbreaks represents a critical chall…
Just Enough Shifts: Mitigating Over-Refusal in Aligned Language Models with Targeted Representation Fine-Tuning
Mahavir Dabas, Si Chen, Charles Fleming +2
Safety alignment is crucial for large language models (LLMs) to resist malicious instructions but often results in over-refusals, where benign prompts are unnecessarily rejected, i…
DiPT: Enhancing LLM reasoning through diversified perspective-taking
Hoang Anh Just, Mahavir Dabas, Lifu Huang +2
Existing work on improving language model reasoning typically explores a single solution path, which can be prone to errors. Inspired by perspective-taking in social studies, this…