Search papers, labs, and topics across Lattice.
Shanghai Jiao Tong University
1
0
2
Quantized reasoning models overthink more, negating efficiency gains, but a new method precisely targets and penalizes these "overthinking" tokens without retraining.
