@aimointerp

Which LLM understands the math, and which is faking it? A competition on robust vs. spurious reasoning in language models. @NeurIPSConf 2026

aimo-interp.github.io
Joined July 2026
  • Tweets8
  • Following9
  • Followers26
3Media
🚀 3,000+ submissions, 300+ teams, $12,500 in prizes. 2 weeks left in the #NeurIPS #AIMOInterp Challenge! Think you understand how AI works? Come show us and claim your award! aimo-interp.github.io
1
1
72
22
Help us build interpretability methods for real-world model monitoring! 🔬 We’re announcing new White-Box Awards in the AIMO Interpretability Challenge: $1,000 per track for methods that look inside the model (activations, patching, ...) Deadline Nov 1 → aimo-interp.github.io
1
2
54
A month left in the AIMO #Interpretability Challenge @NeurIPSConf 🧠 📊 1k+ submissions predicting whether #LLMs reason robustly on Olympiad-level math problems. ⚠️ Remember: high eval acc ≠ high test score. ⏳ Still time until 1 Nov. 👉 aimo-interp.github.io/
1k+ submissions predicting whether #LLMs reason robustly on olympiad-level math problems.

ALT 1k+ submissions predicting whether #LLMs reason robustly on olympiad-level math problems.

2
2
100
Super cool work! Maybe these circuits could also help disentangle robust from spurious reasoning on olympiad-level math problems? The main phase of our @NeurIPSConf competition just started 😉 aimo-interp.github.io
Andrea de Varda@devarda_aSep 7
New preprint! 🤖➕➗ Shared circuits predict whether LLMs generalize across formats in arithmetic reasoning w/ @sanapandey, @pengrui_han, @jacobandreas, and @ev_fedorenko If you can solve 2+5, you can solve "two plus five" without learning math again. LLMs are more brittle (1/5)
1
2
170
Amazing! We'd love insights from open LMs too! "Understand" is the hard part, and our @NeurIPSConf competition. Do the most capable LMs reason robustly on Olympiad-level math problems, or just fake it? 🚨Main phase is on, two months left. aimo-interp.github.io
Sathvik Redrouthu
@_sathvikrSep 4
We partnered with @AnthropicAI @OpenAI to host the world's first math hackathon. 40 hours, $2M compute - solve, understand, and present an open problem. Applications are live: mathathonchallenge.com
1
3
108
📣 The AIMO #Interpretability Challenge at @NeurIPSConf is now in its main phase, with 120 teams already participating. Join us in better understanding whether LMs reason robustly about challenging math problems. Submissions close November 1. aimo-interp.github.io
1
1
46
Do #LLMs understand the hardest math problems—or just fake it? Join the first @NeurIPSConf competition on #Interpretability and #Generalization: AIMO Interpretability Challenge 🏆 $12.5K prizes 📐 Unpublished olympiad problems 🧠 Frontier AIMO models aimo-interp.github.io
Dark blue graphic with faint mathematical formulas in the background. Heading: "Do LLMs understand the hardest math problems—or just fake it?" Below: "The first NeurIPS competition on Interpretability and Generalization", then "AIMO Interpretability Challenge", then "$12.5K prizes · Unpublished olympiad problems · Frontier AIMO models". The URL aimo-interp.github.io appears at the bottom left, with the NeurIPS logo bottom right.

ALT Dark blue graphic with faint mathematical formulas in the background. Heading: "Do LLMs understand the hardest math problems—or just fake it?" Below: "The first NeurIPS competition on Interpretability and Generalization", then "AIMO Interpretability Challenge", then "$12.5K prizes · Unpublished olympiad problems · Frontier AIMO models". The URL aimo-interp.github.io appears at the bottom left, with the NeurIPS logo bottom right.

3
2
10
3,163
Load more