Search papers, labs, and topics across Lattice.
Affiliation:
3
0
5
Small proxy models can reveal near-optimal SFT-RL budget allocations for large LLMs, streamlining the post-training process significantly.
Even with intermittent access, control variates can significantly boost multi-armed bandit performance, offering a practical advantage in real-world scenarios like wireless networks.
LLMs can be made significantly safer by steering their latent space trajectories with Control Barrier Functions, preventing unsafe outputs without retraining.
