14 16 2

Taiwei Shi

MaksimSTW

https://maksimstw.github.io/

AI & ML interests

alignment, human-AI collaboration, and computational social science

Recent Activity

upvoted a paper about 4 hours ago

How Instruction and Reasoning Data shape Post-Training: Data Quality through the Lens of Layer-wise Gradients

upvoted a collection about 6 hours ago

Papers from LIME Lab

new activity 5 days ago

lime-nlp/DeepScaleR_Difficulty:Request for sharing the rollout outputs for success rate calculation

View all activity

Organizations

MaksimSTW's activity

upvoted a paper about 4 hours ago

How Instruction and Reasoning Data shape Post-Training: Data Quality through the Lens of Layer-wise Gradients

Paper • 2504.10766 • Published 1 day ago • 14

upvoted a collection about 6 hours ago

Papers from LIME Lab

Collection

Papers from LIME Lab • 7 items • Updated 7 days ago • 2

upvoted a paper 5 days ago

C3PO: Critical-Layer, Core-Expert, Collaborative Pathway Optimization for Test-Time Expert Re-Mixing

Paper • 2504.07964 • Published 6 days ago • 58

upvoted a paper 6 days ago

Missing Premise exacerbates Overthinking: Are Reasoning Models losing Critical Thinking Skill?

Paper • 2504.06514 • Published 7 days ago • 33

upvoted a paper 7 days ago

Efficient Reinforcement Finetuning via Adaptive Curriculum Learning

Paper • 2504.05520 • Published 8 days ago • 9

upvoted a paper 15 days ago

Discovering Knowledge Deficiencies of Language Models on Massive Knowledge Base

Paper • 2503.23361 • Published 17 days ago • 6

upvoted a collection 17 days ago

Difficulty Estimation Math Datasets

Collection

We perform difficulty estimation on popular math datasets. • 5 items • Updated 7 days ago • 1

upvoted 2 papers about 2 months ago

On the Trustworthiness of Generative Foundation Models: Guideline, Assessment, and Perspective

Paper • 2502.14296 • Published Feb 20 • 46

Detecting and Filtering Unsafe Training Data via Data Attribution

Paper • 2502.11411 • Published Feb 17 • 1

upvoted a paper 3 months ago

WildFeedback: Aligning LLMs With In-situ User Interactions And Feedback

Paper • 2408.15549 • Published Aug 28, 2024 • 1

upvoted 3 papers 5 months ago

How Susceptible are Large Language Models to Ideological Manipulation?

Paper • 2402.11725 • Published Feb 18, 2024 • 1

Can Language Model Moderators Improve the Health of Online Discourse?

Paper • 2311.10781 • Published Nov 16, 2023 • 1

CoAnnotating: Uncertainty-Guided Work Allocation between Human and Large Language Models for Data Annotation

Paper • 2310.15638 • Published Oct 24, 2023 • 1

upvoted a paper 6 months ago

CLIMB: A Benchmark of Clinical Bias in Large Language Models

Paper • 2407.05250 • Published Jul 7, 2024 • 2

upvoted 2 papers about 1 year ago

Safer-Instruct: Aligning Language Models with Automated Preference Data

Paper • 2311.08685 • Published Nov 15, 2023 • 1

TrustLLM: Trustworthiness in Large Language Models

Paper • 2401.05561 • Published Jan 10, 2024 • 70