Lexsi Labs

company

https://lexsi.ai

Activity Feed Request to join this org

AI & ML interests

Frontier research around Safe and aligned intelligence

Recent Activity

Kaushal101 updated a model 2 minutes ago

Lexsi/llama32-3b-legal-sft-drift

Kaushal101 published a model 3 minutes ago

Lexsi/llama32-3b-legal-sft-drift

Kaushal101 updated a model 25 minutes ago

Lexsi/llama32-3b-gsm8k-sft-drift

View all activity

Papers

Forgetting That Sticks: Quantization-Permanent Unlearning via Circuit Attribution

C-ΔΘ: Circuit-Restricted Weight Arithmetic for Selective Refusal

View all Papers

Lexsi 's Papers 9

Submitted by

Pratinav Seth

Forgetting That Sticks: Quantization-Permanent Unlearning via Circuit Attribution

Lexsi

Submitted by

Pratinav Seth

C-ΔΘ: Circuit-Restricted Weight Arithmetic for Selective Refusal

Lexsi

2

Submitted by

Pratinav Seth

Orion-MSP: Multi-Scale Sparse Attention for Tabular In-Context Learning

Lexsi

Submitted by

Pratinav Seth

TabTune: A Unified Library for Inference and Fine-Tuning Tabular Foundation Models

Lexsi

Interpretability as Alignment: Making Internal Understanding a Design Principle

Lexsi

Interpretability-Aware Pruning for Efficient Medical Image Analysis

Lexsi

Bridging the Gap in XAI-Why Reliable Metrics Matter for Explainability and Compliance

Lexsi

xai_evals : A Framework for Evaluating Post-Hoc Local Explanation Methods

Lexsi

DLBacktrace: A Model Agnostic Explainability for any Deep Learning Models

Lexsi