Picture for Daniil Gavrilov

Daniil Gavrilov

Mechanistic Permutability: Match Features Across Layers

Add code
Oct 10, 2024
Viaarxiv icon

Learn Your Reference Model for Real Good Alignment

Add code
Apr 15, 2024
Viaarxiv icon

Linear Transformers with Learnable Kernel Functions are Better In-Context Models

Add code
Feb 16, 2024
Viaarxiv icon

Democratized Diffusion Language Model

Add code
May 18, 2023
Viaarxiv icon

Ahead-of-Time P-Tuning

Add code
May 18, 2023
Viaarxiv icon

Linear Interpolation In Parameter Space is Good Enough for Fine-Tuned Language Models

Add code
Nov 22, 2022
Viaarxiv icon

Classifiers are Better Experts for Controllable Text Generation

Add code
May 15, 2022
Figure 1 for Classifiers are Better Experts for Controllable Text Generation
Figure 2 for Classifiers are Better Experts for Controllable Text Generation
Figure 3 for Classifiers are Better Experts for Controllable Text Generation
Figure 4 for Classifiers are Better Experts for Controllable Text Generation
Viaarxiv icon

PALBERT: Teaching ALBERT to Ponder

Add code
Apr 07, 2022
Figure 1 for PALBERT: Teaching ALBERT to Ponder
Figure 2 for PALBERT: Teaching ALBERT to Ponder
Figure 3 for PALBERT: Teaching ALBERT to Ponder
Figure 4 for PALBERT: Teaching ALBERT to Ponder
Viaarxiv icon

FastRPB: a Scalable Relative Positional Encoding for Long Sequence Tasks

Add code
Feb 23, 2022
Figure 1 for FastRPB: a Scalable Relative Positional Encoding for Long Sequence Tasks
Figure 2 for FastRPB: a Scalable Relative Positional Encoding for Long Sequence Tasks
Figure 3 for FastRPB: a Scalable Relative Positional Encoding for Long Sequence Tasks
Figure 4 for FastRPB: a Scalable Relative Positional Encoding for Long Sequence Tasks
Viaarxiv icon

Implicit Unlikelihood Training: Improving Neural Text Generation with Reinforcement Learning

Add code
Jan 11, 2021
Figure 1 for Implicit Unlikelihood Training: Improving Neural Text Generation with Reinforcement Learning
Figure 2 for Implicit Unlikelihood Training: Improving Neural Text Generation with Reinforcement Learning
Figure 3 for Implicit Unlikelihood Training: Improving Neural Text Generation with Reinforcement Learning
Figure 4 for Implicit Unlikelihood Training: Improving Neural Text Generation with Reinforcement Learning
Viaarxiv icon