5 papers
Harnessing the Potential of Optimizing Data Mixtures via Bayesian Domain Reweighting
Xiang Yuan, Kaiqing Lei, Zhenyu Jin +3
The performance of Large Language Models (LLMs) is fundamentally influenced by the distributional composition of multi-domain pre-training data. While manual heuristics were preval…
Machine Learning-based Separation of the He I 10830Å Chromospheric Signal: Quantitative Analysis of Chromosphere-Corona Intensity in the Quiet Sun
Huaiming Li, Fangyu Xu, Yi Bi +1
The He I 10830Å line, a crucial optically thin chromospheric line, is frequently used to study coronal heating and vertical coupling across the chromosphere-corona interface. Howev…
A Distributional View for Visual Mechanistic Interpretability: KL-Minimal Soft-Constraint Principle
Guancheng Zhou, Yisi Luo, Zhengfu He +5
Most current paradigms in visual mechanistic interpretability (MI) remain confined to interpreting internal units of the vision model via heuristic methods (e.g., top- activatio…
Using Neural Emulators and Hamiltonian Monte Carlo to constrain the Epoch of Reionization's History with the Ly Forest Power Spectrum
Diego González-Hernández, Caitlin Doughty, Molly Wolfson +2
The Lyman-alpha (Ly) forest at offers a primary probe to constrain the history of the Epoch of Reionization (EoR), retaining thermal and ionization signatures imprint…
Neural network emulator to constrain the high- IGM thermal state from Lyman- forest flux auto-correlation function
Zhenyu Jin, Molly Wolfson, Joseph F. Hennawi +1
We present a neural network emulator to constrain the thermal parameters of the intergalactic medium (IGM) at using the Lyman- (L…