影响力指数

92.84/100

前 0.4%

全站排名 #255

发表论文26 篇

平均评分6.0

年均产出8.7 篇/年

Aditi Raghunathan

Assistant Professor@Carnegie Mellon University·美国·OpenReview

研究方向

AI safety · Memorization in LLMs · Understanding limits of foundation models · Robustness to distribution shifts · Fairness · Adversarial examples

Differential Smoothing Mitigates Sharpening and Improves LLM Reasoning

ICLR 2026Rejected

ImpossibleBench: Measuring LLMs' Propensity of Exploiting Test Cases

ICLR 2026Poster

Watch the Weights: Unsupervised monitoring and control of fine-tuned LLMs

ICLR 2026Poster

Mode-conditioning unlocks superior test-time compute scaling

ICLR 2026Poster

Jailbreaking in the Haystack

ICLR 2026Desk Rejected

Roll the dice & look before you leap: Going beyond the creative limits of next-token prediction

Context-Parametric Inversion: Why Instruction Finetuning May Not Actually Improve Context Reliance

Scaling Laws for Precision

Overtrained Language Models Are Harder to Fine-Tune

ICML 2025Poster

Weight ensembling improves reasoning in language models

COLM 2025Poster

Mitigating Modal Imbalance in Multimodal Reasoning

COLM 2025Poster

Reasoning as an Adaptive Defense for Safety

NeurIPS 2025Poster

Dissecting Adversarial Robustness of Multimodal LM Agents

ICLR 2025Poster

Repetition Improves Language Model Embeddings

ICLR 2025Poster

Memorization Sinks: Isolating Memorization during LLM Training

ICML 2025Poster

Testing the Limits of Jailbreaking with the Purple Problem

ICLR 2025Rejected

Lowering Data Diversity can Accelerate Training: Case Studies in Synthetic Tasks

ICLR 2025Rejected

合作者 (20)

Jacob Mitchell Springer