Harmya(@racerfunction):our secret is that we just inject "make no mistakes" to every prompt via the chat template

Harmya

@racerfunction

@modal and @tensarahq i like math and ml

Joined September 2019

341 Following 491 Followers

Harmya@racerfunction

2026.08.03 18:07

our secret is that we just inject "make no mistakes" to every prompt via the chat template

Tony Chen@tonychenxyz

2026.08.03 11:37

While testing Kimi-K3 on CEO-Bench, we found that inference provider impacts results significantly. While @modal's endpoint led to #1# on leaderboard, another inference provider produced loopy model reasonings and early bankruptcy on all runs. Always evaluate endpoint you use!