Lab Practices for advanced learners
687 resources · page 21 of 22
DFM Mimir v1: An Open HRM Delivering Frontier Performance at 1B Parameters
Creator not supplied; see original source.
arXiv paper (2026) describing DFM Mimir v1, a 1B-parameter Danish open model.
DeepSeek-V3 Technical Report
Creator not supplied; see original source.
arXiv paper (2024) DeepSeek-V3 Technical Report: MLA attention, DeepSeekMoE and FP8 training.
deepset (German’s Next Language Model)
Creator not supplied; see original source.
arXiv HTML version of deepset's German's Next Language Model paper (2020) on GBERT and GELECTRA.
Eval awareness in Claude Opus 4.6's BrowseComp performance
Creator not supplied; see original source.
Anthropic engineering post on Claude Opus 4.6 showing evaluation awareness on the BrowseComp benchmark.
ÚFAL at MultiLexNorm 2021: Improving Multilingual Lexical Normalization by Fine-tuning ByT5
Creator not supplied; see original source.
arXiv paper (2021) describing ÚFAL's system for multilingual lexical normalisation at MultiLexNorm 2021.
GLM-4.5 ARC tech report
Creator not supplied; see original source.
arXiv paper (2025) GLM-4.5 technical report on Z.ai's agentic, reasoning and coding foundation models.
Google DeepMind Publications
Creator not supplied; see original source.
Google DeepMind publications index: research papers across AI and science.
Latxa: An Open Language Model and Evaluation Suite for Basque
Creator not supplied; see original source.
arXiv paper (2024) introducing Latxa, an open Basque language model and evaluation suite.
Meltemi: The first open Large Language Model for Greek
Creator not supplied; see original source.
arXiv paper (2024) introducing Meltemi, an open large language model for Greek.
Mitigating Catastrophic Forgetting in Language Transfer via Model Merging
Creator not supplied; see original source.
EMNLP 2024 Findings paper (INSAIT) on reducing catastrophic forgetting in language transfer via Branch-and-Merge model merging.
GePpeTto Carves Italian into a Language Model
Creator not supplied; see original source.
CEUR-WS paper (PDF) presenting GePpeTto, an Italian GPT-2 language model.
jina-embeddings-v4: Universal Embeddings for Multimodal Multilingual Retrieval
Creator not supplied; see original source.
arXiv paper (2025) introducing jina-embeddings-v4, universal multimodal multilingual embeddings.
Swedish Whispers: Leveraging a Massive Speech Corpus for Swedish Speech Recognition
Creator not supplied; see original source.
arXiv paper (2025) by KBLab on Swedish Whisper models trained on a large Swedish speech corpus.
Kimi K3: Open Frontier Intelligence
Creator not supplied; see original source.
arXiv paper (2026) Kimi K3 technical report from Moonshot AI.
Kling-Omni Technical Report
Creator not supplied; see original source.
arXiv HTML version of the Kling-Omni technical report (2025) on Kuaishou's video model architecture and training.
LAION-5B: An open large-scale dataset for training next generation image-text models
Creator not supplied; see original source.
arXiv paper (2022) introducing LAION-5B, an open dataset of 5.85 billion image-text pairs.
EXAONE 4.0
Official model guidance with reasoning and deployment examples.
LoRA Without Regret
Creator not supplied; see original source.
Thinking Machines Lab post LoRA Without Regret: when LoRA matches full fine-tuning, with hyperparameter guidance.
Luma Unified Models (Uni-1)
Creator not supplied; see original source.
Luma AI Uni-1 page: multimodal reasoning model that generates images.
MERA paper
Creator not supplied; see original source.
arXiv paper (2024) introducing MERA, a benchmark for evaluating LLMs in Russian.
MiniMax-M1 paper
Creator not supplied; see original source.
arXiv paper (2025) describing MiniMax-M1, scaling test-time compute with lightning attention.
MiniMax-M2 repo
by MiniMax
Examine MiniMax-M2's model release and agent-oriented evaluation report.
MiniMax Sparse Attention
Creator not supplied; see original source.
arXiv paper (2026) describing MiniMax Sparse Attention, the attention design accompanying MiniMax-M3.
Mistral AI Research
Creator not supplied; see original source.
Mistral AI research page listing papers and technical reports on its language, speech, document and robotics models.
Kimi K2 repo
by Moonshot AI
Read the Kimi K2 model overview and deployment material.
Muon is Scalable for LLM Training (Moonlight)
Creator not supplied; see original source.
arXiv paper (2025) by Moonshot AI showing the Muon optimizer scales to LLM training, with the Moonlight model.
Annotation-Efficient Vision-Language Model Adaptation to the Polish Language Using the LLaVA Framework
Creator not supplied; see original source.
arXiv paper (2026) by NASK on annotation-efficient adaptation of vision-language models to Polish.
CLOVA Tech Blog
Creator not supplied; see original source.
NAVER CLOVA tech blog with technical posts on HyperCLOVA X models and their serving.
NorwAI's Large Language Models: Technical Report
Creator not supplied; see original source.
arXiv paper (2026) technical report on NorwAI's Norwegian large language models.
Nous Research blog
Creator not supplied; see original source.
Nous Research blog with posts on Hermes models, the Psyche training network and research.
GlórIA: A Generative and Open Large Language Model for Portuguese
Creator not supplied; see original source.
arXiv paper (2024) introducing GlórIA, an open generative language model for European Portuguese.
OpenLLM-Ro: Technical Report on Open-source Romanian LLMs
Creator not supplied; see original source.
arXiv paper (2024) technical report on OpenLLM-Ro, open-source Romanian LLMs.