Pere Martra

oopere

AI & ML interests

AI Research Engineer | Structured Pruning & LLM Efficiency | Fairness-Aware Optimization | Author “Rearchitecting LLMs” @Manning | 2K GitHub ⭐

Recent Activity

updated a collection about 1 hour ago

Rearchitecting LLMs

liked a model about 1 hour ago

oopere/SmolLM2-1.7B-ClinicalNER

updated a model about 3 hours ago

oopere/SmolLM2-1.7B-ClinicalNER

View all activity

Organizations

upvoted an article 10 days ago

Article

Fairness Pruning: Finding Demographic Bias Neurons in LLMs (And What Happens When You Remove Them)

oopere

•

19 days ago

• 1

upvoted a collection 2 months ago

Rearchitecting LLMs

Collection

Companion models for Rearchitecting LLMs (Manning). Pruning, Distillation, Fine-Tuning, Attention Optimization & Activation Analysis for efficient SLM • 6 items • Updated about 1 hour ago • 1

upvoted a paper 6 months ago

Fragile Knowledge, Robust Instruction-Following: The Width Pruning Dichotomy in Llama-3.2

Paper • 2512.22671 • Published Dec 27, 2025 • 1

upvoted an article 7 months ago

Article

Making LLMs Smaller Without Breaking Them: A GLU-Aware Pruning Approach

oopere

•

Nov 24, 2024

• 20

upvoted 2 articles 11 months ago

Article

Good answers are not necessarily factual answers: an analysis of hallucination in leading LLMs

davidberenstein1957

•

May 7, 2025

• 42

Article

LLMs recognise bias but also reproduce harmful stereotypes: an analysis of bias in leading LLMs

davidberenstein1957

•

Jul 2, 2025

• 16

upvoted an article about 1 year ago

Article

Welcome Gemma 3: Google's all new multimodal, multilingual, long context open LLM

ariG23498, merve, pcuenq, reach-vb

•

Mar 12, 2025

• 497

upvoted 2 articles over 1 year ago

Article

Training a Gemma 2 2B-IT for Reasoning with GRPO

lmassaron

•

Mar 18, 2025

• 6

Article

Open-R1: a fully open reproduction of DeepSeek-R1

eliebak, lvwerra, lewtun

•

Jan 28, 2025

• 889

upvoted a paper over 1 year ago

Shortened LLaMA: A Simple Depth Pruning for Large Language Models

Paper • 2402.02834 • Published Feb 5, 2024 • 17

upvoted 2 articles almost 2 years ago

Article

Preference Tuning LLMs with Direct Preference Optimization Methods

kashif, edbeeching, lewtun, lvwerra, osanseviero

•

Jan 18, 2024

• 84

Article

How NuminaMath Won the 1st AIMO Progress Prize

yfleureau, liyongsea, edbeeching, lewtun, benlipkin, romansoletskyi, vwxyzjn, kashif

•

Jul 11, 2024

• 128

upvoted a collection about 2 years ago

Preference Datasets for DPO

Collection

This collection contains a list of curated preference datasets for DPO fine-tuning for intent alignment of LLMs • 7 items • Updated Dec 11, 2024 • 49

upvoted 2 articles about 2 years ago

Article

🪆 Introduction to Matryoshka Embedding Models

tomaarsen, Xenova, osanseviero

•

Feb 23, 2024

• 210

Article

Making LLMs even more accessible with bitsandbytes, 4-bit quantization and QLoRA

ybelkada, timdettmers, artidoro, sgugger, smangrul

•

May 24, 2023

• 180

Pere Martra

AI & ML interests

Recent Activity

Organizations

oopere's activity

Fairness Pruning: Finding Demographic Bias Neurons in LLMs (And What Happens When You Remove Them)

Making LLMs Smaller Without Breaking Them: A GLU-Aware Pruning Approach

Good answers are not necessarily factual answers: an analysis of hallucination in leading LLMs

LLMs recognise bias but also reproduce harmful stereotypes: an analysis of bias in leading LLMs

Welcome Gemma 3: Google's all new multimodal, multilingual, long context open LLM

Training a Gemma 2 2B-IT for Reasoning with GRPO

Open-R1: a fully open reproduction of DeepSeek-R1

Preference Tuning LLMs with Direct Preference Optimization Methods

How NuminaMath Won the 1st AIMO Progress Prize

🪆 Introduction to Matryoshka Embedding Models

Making LLMs even more accessible with bitsandbytes, 4-bit quantization and QLoRA