Skip to content
Logarithms
Our businessesHow we operateContact
LOGARITHMSLogarithms Labs home
HomeResearchBenchmarksHuman DataEnvironmentsTasks & Rubrics
Contact→

Research

Papers and research notes.

Technical work from Logarithms Labs, written to explain the problem, method, evidence and limits clearly enough for others to use.

01

Benchmark research · August 2026

TriageBench: Benchmarking Sequential Clinical Triage Reasoning

Why clinical triage needs an evaluation of asking, stopping and escalation before the full case is known.
→
02

Research programme · August 2026

Larrey-1

A research plan for training medical models to ask, stop and escalate during sequential clinical triage.
→
03

Paper · May 2026

Coarse Semantic Injection for LLM-Conditioned Structured Indoor Prediction

A small semantic interface that improves thin-structure prediction in point-cloud language models.
→
Logarithms Labs

Reinforcement learning for critical domains.

© 2026 Logarithms Labs

A Logarithms company ↗

Logarithms

We build and acquire products that democratise access.

Businesses

DecorLMScaffoldRight Click UseOne Human UnicornLogarithms LabsGum Technologies

Explore

Our businessesHow we operateHuman DataResearch paper

Contact

Contact Logarithmshello@logarith.ms
© 2026 Logarithms
Light / Dark