← Back to People
Lead researcher on Olmo 3 model development, discussing the reasoning and RL training improvements.
4
Mentions
3
Articles
2
Outlets
Dec 2025 — Apr 2026
Coverage
Positive
Sentiment
Associated Topics
Companies Covered
Coverage Patterns
How media typically covers Hanna Hajishirzi
Article Types
- announcement1
Media Angles
- technical1
Narrative Framing
- breakthrough1
Appears In
Sentiment Distribution
Based on 3 scored articles
Positive 67%Mixed 33%
Associated AI Models
Olmo 3.1 Think 32B1Olmo 3.1 Instruct 32B1Olmo 3.1 RL Zero 7B Code1Olmo 3.1 RL Zero 7B Math1Olmo 31
Writes About
Appears Alongside
Articles
Most recent firstCited In
Research or work cited
Allen AI Scales Reinforcement Learning Training to 32B Without Changing Model Architecture
Allen AIannouncementpositiveDec 15, 2025
Allen AI released Olmo 3.1 32B models with extended reinforcement learning training achieving 5+ point gains on AIME and 20+ point improvements on IFBench while maintaining full model flow transparency.
“Referenced in "Allen AI Scales Reinforcement Learning Training to 32B Without Changing Model Ar" in Allen AI”
Open Source AIReinforcement LearningFoundation ModelsLLMs
Also Mentioned In
Referenced in coverage
OLMo 3 asynchronous RL
Allen AIpositiveApr 7, 2026
“Lead researcher on Olmo 3 model development, discussing the reasoning and RL training improvements.”
Microsoft AI execuhired AI2 leadership
XmixedMar 25, 2026
“Recruited from Allen Institute for AI to join Microsoft's Superintelligence team as part of leadership acquisition.”