Ai2

non-profit

Verified

https://allenai.org/

AI & ML interests

Building breatkthrough AI to solve the world's biggest problems.

Recent Activity

Ai2Comms published an article about 13 hours ago

FairStress: A deeper evaluation of fairness in LLM decision-making

faezeb published a dataset about 14 hours ago

allenai/fairstress

faezeb updated a dataset 1 day ago

allenai/fairstress

View all activity

Papers

Learning to Solve Hard Problems in RL for LLMs by Never Giving Up

MolmoMotion: Forecasting Point Trajectories in 3D with Language Instruction

View all Papers

Articles

FairStress: A deeper evaluation of fairness in LLM decision-making

about 13 hours ago

HiRO-ACEv2: Simulating extreme weather at 3 km resolution for a fraction of the cost

Impactful scheduling for GPU clusters

Open-sourcing AstaBrief, the fast report-generation model in Asta

Introducing Olmo-core 3: Open, scalable training infrastructure for large MoEs

BenchMIRT: What are LLM benchmarks actually measuring?

Introducing OlmoEarth embeddings: Custom embedding exports from OlmoEarth Studio for downstream analysis

TutorMoments: Do AI tutors know when to help and when to hold back?

The OlmoEarth Platform: Geospatial inference at planetary scale

What building Shippy taught us about building agents

DiScoFormer: One transformer for density and score, across distributions

Which tokens does a hybrid model predict better?

MolmoMotion: Language-guided 3D motion forecasting

olmo-eval: An evaluation workbench for the model development loop

OlmoEarth v1.1: A more efficient family of Earth observation models

EMO: Pretraining mixture of experts for emergent modularity

View all articles

published an article about 13 hours ago

Article

FairStress: A deeper evaluation of fairness in LLM decision-making

allenai

•

about 13 hours ago

• 5

published a dataset about 14 hours ago

allenai/fairstress

Updated 1 day ago • 3

updated a dataset 1 day ago

allenai/fairstress

Updated 1 day ago • 3

updated a dataset 2 days ago

allenai/asta-summary-citation-counts

Viewer • Updated 2 days ago • 127M • 1.34k • 12

updated a dataset 6 days ago

allenai/asta-bench-submissions

Updated 6 days ago • 87 • 3

updated a collection 6 days ago

ACE

Ai2 Climate Emulator (ACE) is a family of fast ML models that simulate weather and climate variability over time scales from hours to centuries. • 15 items • Updated 6 days ago • 18

published an article 7 days ago

Article

HiRO-ACEv2: Simulating extreme weather at 3 km resolution for a fraction of the cost

allenai

•

7 days ago

• 14

published a model 7 days ago

allenai/HiRO-ACEv2

Updated 7 days ago • 5

updated a model 7 days ago

allenai/HiRO-ACEv2

Updated 7 days ago • 5

submitted a paper to Daily Papers 7 days ago

Scaling Repeated-Data Pretraining: Batch Size, Regularization, and GPU-Time Tradeoffs

Paper • 2610.15965 • Published 10 days ago • 4

updated a bucket 7 days ago

allenai/ai2cm-xshield

in allenai/Olmo-3.1-32B-Think 7 days ago

Add LEXam-hard evaluation result

#6 opened about 1 month ago by

published an article 12 days ago

Article

Impactful scheduling for GPU clusters

allenai

•

12 days ago

• 54

published a bucket 12 days ago

allenai/ai2cm-xshield

authored a paper 13 days ago

Old Ideas, Novel Problems: The Instability of LLM-Based Novelty Evaluation

Paper • 2610.02022 • Published 20 days ago

authored 4 papers about 1 month ago

Operads for compositional reasoning in LLMs

Paper • 2606.13634 • Published Jun 11

From Token Probabilities to Semantic Constraints: Towards Declarative Probabilistic Evaluation of Language Models

Paper • 2609.13520 • Published Sep 11

Operadic consistency: a label-free signal for compositional reasoning failures in LLMs

Paper • 2606.13649 • Published Jun 11

ArtifactLinker: Linking Scientific Artifacts for Automatic State-of-the-Art Discovery

Paper • 2605.16902 • Published May 16 • 1

submitted a paper to Daily Papers about 1 month ago

Learning to Solve Hard Problems in RL for LLMs by Never Giving Up

Paper • 2609.13443 • Published Sep 11 • 14