Articles
Discover actionable blogs and thought leadership on Explainable AI, Responsible AI, AI governance, Machine Learning, MLOps, ML observability, and model monitoring.
Featured Articles
All Articles
The Lexsi Alignment and Safety Stack: correcting Gemma-4-31B for multilingual and cybersecurity safety, in the weights Copy
October 8, 2026
AI Alignment
AI Safety
Lexsi Labs, Mumbai
Lexsi Labs, Paris
Trustworthy AI
AI risk management
AI Safety
Open-sourcing the Lexsi Alignment and Safety Stack: frontier safety research for auditing, correcting and verifying model weights
September 16, 2026
AI risk management
AI Alignment
Trustworthy AI
TabTune
Debugging TabFM: the story the leaderboard hides
July 20, 2026
Tabular Foundational Models
Fune-tuning
TabTune
Now Shipping: TabTune Regression for Tabular Foundational Models
February 26, 2026
Tabular Foundational Models
Regression
TabTune
TabTune Resampling: Context Engineering for Zero-Shot Imbalanced Tabular Learning
February 25, 2026
In Context Learning
Orion-MSP
Orion-BiX
Lexsi Labs, Mumbai
Lexsi Labs, Paris
AlignTune
Retail Banking Case Study: Why Specialists Beat Generalists on Template-Strict Workflows
February 23, 2026
Notebook
Post training alignment
Lexsi Labs, Mumbai
AlignTune
The Specialization Dividend: Aligning a 4B Model for Wealth Management using AlignTune
February 23, 2026
Post training alignment
Notebook
Lexsi Labs, Mumbai
AlignTune
Open-Sourcing AlignTune: A Unified Interface for Post-Training Alignment
February 23, 2026
Post training alignment
Reinforcement Learning
Fune-tuning
Lexsi Labs, Mumbai
Lexsi Labs, Paris
Reinforcement Learning
Not All Tokens Matter Equally: Teaching RL Which Parts of Reasoning Actually Count
February 13, 2026
Reasoning Models
Post training alignment
Reinforcement Learning
Beyond KL Divergence: Why Your Policy Optimization Might Be Using the Wrong Regularizer
February 12, 2026
Policy Optimization
Reasoning Models
Post training alignment
Mirror Maps
Mechanistic Interpretability
Circuit-Restricted Weight Arithmetic (C-ΔΘ): Offline Safety Without Runtime Cost
February 11, 2026
Lexsi Labs, Mumbai
Mechanistic Interpretability
AI Safety
CiruitKit
No results found.