Search papers, labs, and topics across Lattice.
3
0
7
Interpolating latent representations before decoding yields a reconstruction FID (iFID) that finally aligns with the generation FID of latent diffusion models, achieving ~0.85 correlation where standard rFID fails.
By modeling the distribution of confidence scores, DistriVoting significantly boosts the accuracy of large reasoning models, outperforming existing confidence-based selection methods across diverse benchmarks.
Forget difficulty-based heuristics: InSight leverages weighted mutual information to select RL training data, boosting LLM reasoning and alignment with up to 2.2x speedup.
