Pan Lu

@lupantech

Postdoc @Stanford @stanfordnlp | PhD @CS_UCLA @uclanlp Amazon/Bloomberg/Qualcomm Fellows | Ex @Tsinghua_Uni @Microsoft | Eubiota/AgentFlow/MathVista/ScienceQA

Palo Alto

lupantech.github.io

Replies Reposts Media

Pinned
Pan Lu
@lupantech
Jun 9
Excited to see Agents' Last Exam (ALE) out! agents-last-exam.org As AI agents move toward real-world work, we need rigorous benchmarks to measure their capabilities, limitations, and broader societal and labor-market impact. ALE is an important step toward grounding the
Yiyou Sun
@YiyouSun
Jun 9
“AI agents will outperform humans at almost all jobs by 2026–2027.” - The forecast is everywhere. So we built the exam to test that claim, on real labor-market aligned work. On the hardest tier, top agents pass 2.6%. Meet Agents' Last Exam (ALE), a rolling benchmark measuring
00:00
3
Pan Lu
@lupantech
Jul 5
OctoTools is an ACL 2026 Oral (Top 3.4%)! 🎉 It provides a training-free agent harness for complex reasoning: plug in tools, route tasks, and compose tool-use workflows without extra model training. 🐙🛠️ Since release: 1.5K GitHub stars and 100+ cites! Huge thanks to
Pan Lu
@lupantech
Apr 6
Excited to share that OctoTools has been accepted to ACL 2026. 🐙 OctoTools is our training-free, extensible framework for tool-using agents on complex reasoning tasks. Grateful to the broader community for the support. Our GitHub repo has now reached 1.4K stars. 📣 Huge
2
Pan Lu
@lupantech
Jul 3
Thanks Sohyeon! Super helpful!
Sohyeon Kim
@ohmyksh
Jun 26
If you’re coming to #ICML2026 🇰🇷, this Seoul Guide might be All You Need! I’ve lived in Seoul for 20+ years and curated this all-in-one guide: Top 5 must-visits, 100+ recs with Google Maps, practical local tips, and more. Check it out: ohmyksh.github.io/icml2026/index… DMs are open and
1
Pan Lu
@lupantech
Jun 24
Congrats to Jiaru and team on #LatentMAS being accepted to #ICML2026 as a Spotlight! Cool work pushing multi-agent collaboration beyond text into latent-space reasoning and communication, with strong gains in accuracy, speed, and token efficiency.
Jiaru "Rubin" Zou
@Jiaru_Zou
Jun 24
Excited to share that #LatentMAS has been accepted to ICML 2026 as a spotlight! 💻Code: github.com/Gen-Verse/Late… 📄Paper: arxiv.org/abs/2511.20639 We push multi-agent collaboration into the latent space — beyond human language. Most multi-agent systems rely on text: agents
2
Pan Lu
@lupantech
May 19
Congrats @haotian_yeee and team on SimpleTES and on discovering 21 SOTA solutions across 6 scientific problems! 🚀🔬 What I find especially exciting is the shift from scaling generation to scaling the full scientific discovery loop: propose → evaluate → refine → repeat. 🔁 By
Haotian Ye
@haotian_yeee
May 19
🚀 Today, we’re excited to introduce SimpleTES for scaling the scientific discovery loop. 🧵 I always ask myself: what are we actually scaling in scientific discovery? Most LLM discovery methods focus on test-time scaling generation — more tokens, more agents, more turns. But
1