Chi Jin
@chijinML
Followers
4K
Following
350
Media
19
Statuses
111
Assistant Prof @Princeton. Previously: ML theory, RL & optimization. Now: AI for math, games & decision making.
Princeton, NJ
Joined November 2012
Really excited to announce this LLM formal math prover project --- Goedel, where we achieve the SOTA open-source performance! Kudos to the team!.
🚀 Introducing Goedel-Prover: A 7B LLM achieving SOTA open-source performance in automated theorem proving! 🔥. ✅ Improving +7% over previous open source SOTA on miniF2F.🏆 Ranking 1st on the PutnamBench Leaderboard.🤖 Solving 1.9X total problems compared to prior works on Lean
6
3
76
Congratulations to my brilliant and accomplished student Qinghua Liu @qinghual2020 on his graduation! 🎓 Excited for his next chapter at #OpenAI. Also honored to hood our amazing Princeton graduates: Chia-Hao Li, Kurtland Chua, Kexin Jin, and Yixiao Chen!
2
3
60
RT @sethkarten: Excited to share that the PokeAgent challenge was accepted as a @NeurIPSConf competition!. This should serve as an excellen….
0
5
0
Huge thanks to all the wonderful collaborators @sethkarten @__jakegrigsby__ @steph_milani @kiranvodrahalli @yayitsamyzhang @fangf07 @yukez.
1
0
7
Really excited to share that our Pokémon competition is accepted at NeurIPS 2025!🔥 Get ready for an epic AI challenge featuring both speedruns🎮 and multi-player Pokémon battles⚔️---more details coming soon!
7
34
340
Tested OpenAI o3 on the final exam from my sophomore-level signal processing/information theory course with ~100 students at Princeton 📊 It scored an impressive 89.5—well above the class average of 72.9, but still short of the top student’s 98+ 💯🎯.
10
4
170
🚀 Our new work from the Goedel-prover team at Princeton: benchmarking compositional reasoning on inequality problems in formal math!.
🚨 Easy math, epic fail! 🚨. Our new benchmark, Ineq-Comp, gives formal theorem provers Lean inequalities. then makes tiny tweaks (duplicating variables, squaring terms) that humans handle easily. Most provers collapse. Simple composition is still surprisingly hard!
0
0
19
Writing math proof in Lean is surprisingly addictive. Watching Terence Tao formalize Lean proofs feels like watching a top-tier gamer playing on Twitch. :-)
1
16
145
Wrapped up the semester with a wonderful group dinner 🍽️. Congrats to our amazing undergrads on graduating or completing their independent work! .Honored to have advised my first big round of undergraduate theses/IWs---grateful for the energy, ideas, and hard work💫 #Princeton
1
2
79
RT @GlobalGoalsAI: Day 2 at #MLxFundamentals is ON!. 📚 From theory to practice – Day 2 is all about depth and application!.✔️ Optimisation….
0
1
0
Three papers accepted at ICML on informal math, game theory and Pokémon!
3
4
206
Congrats on the amazing work! Tackling the last 10–20% on miniF2F takes a major leap in capability 🎉. The competition in formal math is getting so intense --- hard to believe Kimina held SOTA for only two weeks before being solidly surpassed by Deepseek.
We just released DeepSeek-Prover V2. - Solves nearly 90% of miniF2F problems.- Significantly improves the SoTA performance on the PutnamBench.- Achieves a non-trivial pass rate on AIME 24 & 25 problems in their formal version. Github:
0
0
17
RT @PrincetonPLI: We're so proud that Princeton researchers have received 1 outstanding paper award and 1 honorable mention at @iclr_conf….
0
33
0
RT @qinghual2020: Thanks so much, Chi! I’m really grateful to have had you as my advisor. I’ll always miss our time working together at Pri….
0
1
0
Better late than never --- @qinghual2020 finally defended his FPO and officially graduated from our group! 🎓 So proud of you, Dr. Liu. You've done so many incredible work on the foundations of reinforcement learning.
2
3
130
RT @ElliotGlazer: Here is Epoch's second AI+Math Chat, feat. Kevin Buzzard, Yang-Hui He, @AlexKontorovich, and.@KristinLauter! Watch as we….
0
8
0
RT @GoogleDeepMind: Human generated data has fueled incredible AI progress, but what comes next? 📈. On the latest episode of our podcast, @….
0
258
0