vicgalle (Victor Gallego)

upvoted a paper 1 day ago

TinyStyler: Efficient Few-Shot Text Style Transfer with Authorship Embeddings

Paper • 2406.15586 • Published 5 days ago • 2

upvoted a paper 4 days ago

Model Merging and Safety Alignment: One Bad Model Spoils the Bunch

Paper • 2406.14563 • Published 6 days ago • 26

upvoted 2 papers 5 days ago

nabla^2DFT: A Universal Quantum Chemistry Dataset of Drug-Like Molecules and a Benchmark for Neural Network Potentials

Paper • 2406.14347 • Published 6 days ago • 93

τ-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains

Paper • 2406.12045 • Published 9 days ago • 4

upvoted a collection 14 days ago

Performant Models

Collection

My top performing models based off my own testing, user feedback and benches. This are the models I would recommend you using. • 8 items • Updated 11 days ago • 18

upvoted a paper 14 days ago

Merging Improves Self-Critique Against Jailbreak Attacks

Paper • 2406.07188 • Published 15 days ago • 3

upvoted a collection 15 days ago

Multimodal Models 🔀

Collection

A collection of multimodal models developed by the Komorebi AI team • 2 items • Updated 8 days ago • 2

upvoted a paper 22 days ago

Aligning to Thousands of Preferences via System Message Generalization

Paper • 2405.17977 • Published 29 days ago • 6

upvoted an article 22 days ago

Article

Uncensor any LLM with abliteration

By

•

13 days ago

• 223

upvoted an article 23 days ago

Article

Expanding Model Context and Creating Chat Models with a Single Click

By

•

Apr 28

• 35

upvoted an article about 2 months ago

Article

Can we create pedagogically valuable multi-turn synthetic datasets from Cosmopedia?

By

•

May 7

• 7

upvoted a paper about 2 months ago

Replacing Judges with Juries: Evaluating LLM Generations with a Panel of Diverse Models

Paper • 2404.18796 • Published Apr 29 • 67

upvoted an article 2 months ago

Article

🦙⚗️ Using Llama3 and distilabel to build fine-tuning datasets

By

•

22 days ago

• 63

upvoted a collection 3 months ago

Configurable Safety Tuning ⚙️

Collection

CST allows for configurable inference-time control of LLM safety levels, so users can dictate model behavior based on the system prompt • 7 items • Updated 13 days ago • 2

upvoted 3 papers 3 months ago

upvoted a collection 3 months ago

Quantized Models (GGUF, IQ, Imatrix)

Collection

Various quantizations of models in the GGUF format. Models with a "checkmark" are personal favorites. An "orange arrow" means it's being uploaded. • 81 items • Updated 3 days ago • 39

upvoted 4 papers 4 months ago

GRATH: Gradual Self-Truthifying for Large Language Models

Paper • 2401.12292 • Published Jan 22 • 2

A Survey on Data Selection for LLM Instruction Tuning

Paper • 2402.05123 • Published Feb 4 • 3

Refined Direct Preference Optimization with Synthetic Data for Behavioral Alignment of LLMs

Paper • 2402.08005 • Published Feb 12 • 1

Suppressing Pink Elephants with Direct Principle Feedback

Paper • 2402.07896 • Published Feb 12 • 8

upvoted a paper 5 months ago

Vision Superalignment: Weak-to-Strong Generalization for Vision Foundation Models

Paper • 2402.03749 • Published Feb 6 • 9

upvoted 3 collections 5 months ago

🛰️🌍 Geospatial Datasets

Collection

A curated collections of diverse geospatial and satellite imagery datasets. • 54 items • Updated Mar 6 • 11

Exotic Frankenmerges 🥨

Collection

Merges of models of different architectures and sizes that end up working surprisingly well • 1 item • Updated 13 days ago • 1

Usable Models

Collection

5 items • Updated Dec 10, 2023 • 2

upvoted 2 collections 6 months ago

Upscaled Models ⏫

Collection

A collection of my frankenmerges, upscaling several models. All of them have the corresponding GGUF variants. • 4 items • Updated 13 days ago • 2

Model Merging

Collection

Model Merging is a very popular technique nowadays in LLM. Here is a chronological list of papers on the space that will help you get started with it! • 30 items • Updated 14 days ago • 187

upvoted a paper 7 months ago

Distilled Self-Critique of LLMs with Synthetic Data: a Bayesian Perspective

Paper • 2312.01957 • Published Dec 4, 2023 • 1

upvoted a collection 8 months ago

Optimised Translation Models 🌍

Collection

A collection of optimised and quantised multilingual translation models • 6 items • Updated Nov 7, 2023 • 3

upvoted a paper 10 months ago

Fast Adaptation with Bradley-Terry Preference Models in Text-To-Image Classification and Generation

Paper • 2308.07929 • Published Jul 15, 2023 • 1

upvoted 3 papers 11 months ago

Personalizing Text-to-Image Generation via Aesthetic Gradients

Paper • 2209.12330 • Published Sep 25, 2022 • 1

ZYN: Zero-Shot Reward Models with Yes-No Questions

Paper • 2308.06385 • Published Aug 11, 2023 • 1

LoraHub: Efficient Cross-Task Generalization via Dynamic LoRA Composition

Paper • 2307.13269 • Published Jul 25, 2023 • 30

Victor Gallego

AI & ML interests

Organizations

vicgalle's activity

TinyStyler: Efficient Few-Shot Text Style Transfer with Authorship Embeddings

Model Merging and Safety Alignment: One Bad Model Spoils the Bunch

nabla^2DFT: A Universal Quantum Chemistry Dataset of Drug-Like Molecules and a Benchmark for Neural Network Potentials

τ-bench: A Benchmark for Tool-Agent-User Interaction in Real-World Domains

Performant Models

Merging Improves Self-Critique Against Jailbreak Attacks

Multimodal Models 🔀

Aligning to Thousands of Preferences via System Message Generalization

Uncensor any LLM with abliteration

Expanding Model Context and Creating Chat Models with a Single Click

Can we create pedagogically valuable multi-turn synthetic datasets from Cosmopedia?

Replacing Judges with Juries: Evaluating LLM Generations with a Panel of Diverse Models

🦙⚗️ Using Llama3 and distilabel to build fine-tuning datasets

Configurable Safety Tuning ⚙️

Configurable Safety Tuning of Language Models with Synthetic Preference Data

Evolutionary Optimization of Model Merging Recipes

Large Language Models are Null-Shot Learners

Quantized Models (GGUF, IQ, Imatrix)

GRATH: Gradual Self-Truthifying for Large Language Models

A Survey on Data Selection for LLM Instruction Tuning

Refined Direct Preference Optimization with Synthetic Data for Behavioral Alignment of LLMs

Suppressing Pink Elephants with Direct Principle Feedback

Vision Superalignment: Weak-to-Strong Generalization for Vision Foundation Models

🛰️🌍 Geospatial Datasets

Exotic Frankenmerges 🥨

Usable Models

Upscaled Models ⏫

Model Merging

Distilled Self-Critique of LLMs with Synthetic Data: a Bayesian Perspective

Optimised Translation Models 🌍

Fast Adaptation with Bradley-Terry Preference Models in Text-To-Image Classification and Generation

Personalizing Text-to-Image Generation via Aesthetic Gradients

ZYN: Zero-Shot Reward Models with Yes-No Questions

LoraHub: Efficient Cross-Task Generalization via Dynamic LoRA Composition