Tomasz Korbak
10 papers · 2021–2024 · 4 venues · across top CS/AI venues
Authors are matched by name, so researchers who share a name ("Yang Liu", "Wei Wang") appear as a single profile and their paper counts, venues and achievements are combined. How identity is handled.
Achievements
Jump to papers ↓+1 more ↓ Show less ↑
🌍 Conference Polyglot · Bronze 🌉 Interdisciplinary Bridge · Bronze 🗺️ Taxonomy Completionist (30) 🐣 Hot Topic Early Bird 👥 Mega-Team · Squad
💎
Century Club · Bronze
Venues
Top co-authors
Research topics
Artificial Intelligence
(10)
Core AI
(8)
Machine Learning
(7)
Natural Language Processing
(7)
Learning Types · Machine Learning
(5)
Deep Learning
(4)
Optimization & Theory
(4)
Generation
(3)
Learning Types · Deep Learning
(3)
Resources & Methods
(3)
Interdisciplinary
(1)
Reinforcement Learning
(1)
Bayesian & Probabilistic · Artificial Intelligence
(1)
Bayesian & Probabilistic · Machine Learning
(1)
Core Methods
(1)
Keywords
language model
(4)
reinforcement learning
(3)
reinforcement learning from human feedback
(3)
kl divergence
(2)
large language model
(2)
language model alignment
(2)
human preference
(2)
distribution matching
(2)
language model fine-tuning
(2)
preference model
(2)
policy gradient
(2)
logistic regression
(1)
conditional generation
(1)
signaling games
(1)
bayesian inference
(1)
imitation learning
(1)
preference alignment
(1)
generative model
(1)
self-supervised learning
(1)
transfer learning
(1)
Papers
Compositional Preference Models for Aligning LMs
ICLR 2024
Many-shot Jailbreaking
NeurIPS 2024
The Reversal Curse: LLMs trained on “A is B” fail to learn “B is A”
ICLR 2024
Towards Understanding Sycophancy in Language Models
ICLR 2024
Aligning Language Models with Preferences through f-divergence Minimization
ICML 2023
Pretraining Language Models with Human Preferences
ICML 2023
Controlling Conditional Language Models without Catastrophic Forgetting
ICML 2022
On Reinforcement Learning and Distribution Matching for Fine-Tuning Language Models with no Catastrophic Forgetting
NeurIPS 2022
RL with KL penalties is better viewed as Bayesian inference
EMNLP 2022
Catalytic Role Of Noise And Necessity Of Inductive Biases In The Emergence Of Compositional Communication
NeurIPS 2021