Researcher — paper.dinq.me

2026 顶会论文

作者位次 #4/7ICLR 2026Poster· Deep Learning / Generative Models and Autoencoders

Rote Learning Considered Useful: Generalizing over Memorized Data in LLMs

Qinyuan Wu · Soumi Das · Mahsa Amani · Bishwamittra Ghosh · Mohammad Aflah Khan · Krishna Gummadi · Muhammad Bilal Zafar

理论与学习动力学OpenReview

作者位次 #4/8ICLR 2026Poster· Social Aspects / Trustworthy Machine Learning

In Agents We Trust, but Who Do Agents Trust? Latent Source Preferences Steer LLM Generations

Mohammad Aflah Khan · Mahsa Amani · Soumi Das · Bishwamittra Ghosh · Qinyuan Wu · Krishna Gummadi · Manish Gupta · Abhilasha Ravichander

智能体与工具对齐与偏好社会与人本OpenReview

一作ACL 2026Poster· Natural Language Processing

Fine-tuning vs. In-context Learning in Large Language Models: A Formal Language Learning Perspective

Bishwamittra Ghosh · Soumi Das · Till Speicher · Qinyuan Wu · Mohammad Aflah Khan · Deepak Garg · Krishna P. Gummadi · Evimaria Terzi

摘要

Large language models (LLMs) operate in two fundamental learning modes – fine-tuning (FT) and in-context learning (ICL) – raising key questions about which mode yields greater language proficiency and whether they differ in their inductive biases. Prior studies comparing FT and ICL have yielded m...

评测与基准理论与学习动力学强化学习与后训练OpenReview 项目页

Bishwamittra Ghosh

获取完整研究档案

联系方式

2026 顶会论文

Rote Learning Considered Useful: Generalizing over Memorized Data in LLMs

In Agents We Trust, but Who Do Agents Trust? Latent Source Preferences Steer LLM Generations

Fine-tuning vs. In-context Learning in Large Language Models: A Formal Language Learning Perspective

常见合作者