Vol-4258⫷ Vol-4259 ⫸Vol-4260
urn:nbn:de:0074-4259-0


Vol-4259/paper12⫷Vol-4259/paper13⫸Vol-4259/paper14
Pablo Pérez-LázaroRocío Aznar-GimenoFrancisco José Lacueva-PérezPaula Peña-LarenaRafael del-Hoyo-Alonso

Delayed Convergence and Emergent CoT Reliance in RL-Tuned Language Models