Picture for Sanjeev Arora

Sanjeev Arora

Metacognitive Capabilities of LLMs: An Exploration in Mathematical Problem Solving

May 20, 2024
Viaarxiv icon

Keeping LLMs Aligned After Fine-tuning: The Crucial Role of Prompt Templates

Add code
Feb 28, 2024
Viaarxiv icon

LESS: Selecting Influential Data for Targeted Instruction Tuning

Add code
Feb 20, 2024
Figure 1 for LESS: Selecting Influential Data for Targeted Instruction Tuning
Figure 2 for LESS: Selecting Influential Data for Targeted Instruction Tuning
Figure 3 for LESS: Selecting Influential Data for Targeted Instruction Tuning
Figure 4 for LESS: Selecting Influential Data for Targeted Instruction Tuning
Viaarxiv icon

Language Models as Science Tutors

Add code
Feb 16, 2024
Figure 1 for Language Models as Science Tutors
Figure 2 for Language Models as Science Tutors
Figure 3 for Language Models as Science Tutors
Figure 4 for Language Models as Science Tutors
Viaarxiv icon

Unlearning via Sparse Representations

Nov 26, 2023
Figure 1 for Unlearning via Sparse Representations
Figure 2 for Unlearning via Sparse Representations
Figure 3 for Unlearning via Sparse Representations
Figure 4 for Unlearning via Sparse Representations
Viaarxiv icon

Skill-Mix: a Flexible and Expandable Family of Evaluations for AI models

Add code
Oct 26, 2023
Figure 1 for Skill-Mix: a Flexible and Expandable Family of Evaluations for AI models
Figure 2 for Skill-Mix: a Flexible and Expandable Family of Evaluations for AI models
Figure 3 for Skill-Mix: a Flexible and Expandable Family of Evaluations for AI models
Figure 4 for Skill-Mix: a Flexible and Expandable Family of Evaluations for AI models
Viaarxiv icon

A Quadratic Synchronization Rule for Distributed Deep Learning

Add code
Oct 22, 2023
Viaarxiv icon

A Theory for Emergence of Complex Skills in Language Models

Jul 29, 2023
Figure 1 for A Theory for Emergence of Complex Skills in Language Models
Figure 2 for A Theory for Emergence of Complex Skills in Language Models
Viaarxiv icon

Trainable Transformer in Transformer

Add code
Jul 03, 2023
Viaarxiv icon

Fine-Tuning Language Models with Just Forward Passes

Add code
May 27, 2023
Figure 1 for Fine-Tuning Language Models with Just Forward Passes
Figure 2 for Fine-Tuning Language Models with Just Forward Passes
Figure 3 for Fine-Tuning Language Models with Just Forward Passes
Figure 4 for Fine-Tuning Language Models with Just Forward Passes
Viaarxiv icon