Search papers, labs, and topics across Lattice.
1
0
2
Freezing pretrained embeddings in a Tiny Recursive Model slashes trainable parameters by 37x while matching or exceeding the quality estimation performance of fine-tuned models.
