Your personalresearch library
Discover papers, track what you're reading, rate and review.Build a library that grows with your curiosity.
Currently Reading
Want to Read
All Papers
Figuring Out Gas & Galaxies In Enzo (FOGGIE) XI: Circumgalactic O VI Emission Traces Clumpy Inflowing Recycled Gas
Cassandra Lochhaas, Molly S. Peeples, Brian W. O'Shea, Jason Tumlinson, Lauren Corlies, Vida Saeedzadeh, Nicolas Lehner, Anna C. Wright, Jessica K. Werk, Cameron W. Trapp, Ramona Augustin, Ayan Acharyya, Britton D. Smith, Carlos J. Vargas
arXiv·2025
The circumgalactic medium (CGM) is host to gas flows into and out of galaxies and regulates galaxy growth, but the multiphase, diffuse gas in this region is challenging to observe. We investigate the properties of gas giving rise to O VI emission from the CGM that upcoming missions, such as the Aspera SmallSat, will be able to map in local galaxies. We use the FOGGIE simulations to predict the O VI emission from edge-on galaxies across the redshift range $z=1\rightarrow0$. O VI emission is brightest surrounding small, clumpy structures near the galaxy where the gas density is high. Most of the O VI surface brightness originates from collisionally ionized, $T\sim10^{5.5}$ K, inflowing gas and is not preferentially aligned with the major or minor axis of the galaxy disk. Simulated galaxies with higher halo masses, higher median CGM gas density, and higher star formation rates produce brighter and more widespread O VI emission in their CGM. We show that while O VI emission primarily originates in inflowing gas, turning off outflows in a simulation without star formation feedback eliminates most of the O VI emission. Enrichment from feedback is necessary to mix with the inflowing gas and allow it to glow in O VI. Collectively, our findings point towards a picture where O VI emission traces warm, ionized envelopes of cooler clouds that are accreting onto the galaxy in a metal-enriched galactic fountain. Finally, we show that the detection limit of Aspera is sufficient to detect O VI emission tens of kpc from the galaxy center for $\sim L^\star$ galaxies.
No ratings yet
View paper →Gaperon: A Peppered English-French Generative Language Model Suite
Nathan Godey, Wissam Antoun, Rian Touchent, Rachel Bawden, Éric de la Clergerie, Benoît Sagot, Djamé Seddah
arXiv·2025
We release Gaperon, a fully open suite of French-English-coding language models designed to advance transparency and reproducibility in large-scale model training. The Gaperon family includes 1.5B, 8B, and 24B parameter models trained on 2-4 trillion tokens, released with all elements of the training pipeline: French and English datasets filtered with a neural quality classifier, an efficient data curation and training framework, and hundreds of intermediate checkpoints. Through this work, we study how data filtering and contamination interact to shape both benchmark and generative performance. We find that filtering for linguistic quality enhances text fluency and coherence but yields subpar benchmark results, and that late deliberate contamination -- continuing training on data mixes that include test sets -- recovers competitive scores while only reasonably harming generation quality. We discuss how usual neural filtering can unintentionally amplify benchmark leakage. To support further research, we also introduce harmless data poisoning during pretraining, providing a realistic testbed for safety studies. By openly releasing all models, datasets, code, and checkpoints, Gaperon establishes a reproducible foundation for exploring the trade-offs between data curation, evaluation, safety, and openness in multilingual language model development.
No ratings yet
View paper →