1

Fine-tuned checkpoints and adapters now fill public repositories, and the most common operation applied to these artifacts is model …

Math and science reasoning benchmarks rely on pass@k, the fraction of sampled chains that reach gold, as the canonical per-example …

We train a contrastive learning music model for stem retrieval, it achieves state-of-the-art due to its phase-aware architecture.

The Platonic Representation Hypothesis suggests that independently trained neural networks converge to increasingly similar latent …

Model merging combines knowledge from separately fine-tuned models, yet success factors remain poorly understood. While recent work …

We show that robustness to post-training quantization (PTQ) is a transferable direction in weight space. We call this direction the …

Neural networks transform high-dimensional data into compact, structured representations, often modeled as elements of a lower …

Transformer components such as non-linear activations and normalization are inherently non-injective, suggesting that different inputs …

CLIP is a discriminative model trained to align images and text in a shared embedding space. Due to its multimodal structure, it serves …

Model merging has recently emerged as a lightweight alternative to ensembling, combining multiple fine-tuned models into a single set …