arXiv subjects

Cheng Cheng

Publications and source records attributed to Cheng Cheng.

At least 127 records · Page 7Linked to original sources

arXiv2024

A strong He II $λ$1640 emitter with extremely blue UV spectral slope at $z=8.16$: presence of Pop III stars?

Cosmic hydrogen reionization and cosmic production of first metals are major phase transitions of the universe occurring during the first billion years after the Big Bang, however these are still underexplored observationally. Using the JWST NIRSpec prism spectroscopy, we report the discovery of a sub-$L_\ast$ galaxy at $z_{\rm spec}=8.1623\pm0.0007$, dubbed RXJ2129-z8HeII, via the detection of a series of strong rest-frame UV/optical nebular emission lines and the clear Lyman break. RXJ2129-z8HeII shows a pronounced UV continuum with an extremely steep (i.e. blue) spectral slope of $β=-2.53_{-0.07}^{+0.06}$, the steepest amongst all spectroscopically confirmed galaxies at $z_{\rm spec}\gtrsim7$, in support of its very hard ionizing spectrum that could lead to a significant leakage of its ionizing flux. Therefore, RXJ2129-z8HeII is representative of the key galaxy population driving the cosmic reionization. More importantly, we detect a strong He II $λ$1640 emission line in its spectrum, one of the highest redshifts at which such a line is robustly detected. Its high rest-frame equivalent width (${\rm EW}=21\pm4$ Angstrom) and extreme flux ratios with respect to UV metal and Balmer lines raise the possibility that part of RXJ2129-z8HeII's stellar populations could be Pop III-like. Through careful photoionization modeling, we show that the physically calibrated phenomenological models of the ionizing spectra of Pop III stars with strong mass loss can successfully reproduce the emission line flux ratios observed in RXJ2129-z8HeII. Assuming the Eddington limit, the total mass of the Pop III stars within this system is estimated to be $7.8\pm1.4\times10^5 M_\odot$. To date, this galaxy presents the most compelling case in the early universe where trace Pop III stars might coexist with metal-enriched populations.

astro-ph.GA↗

arXiv2024

Birds of a Feather: Resolving Stellar Mass Assembly With JWST/NIRCam in a Pair of Kindred $z \sim 2$ Dusty Star-forming Galaxies Lensed by the PLCK G165.7+67.0 Cluster

We present a new parametric lens model for the G165.7+67.0 galaxy cluster, which was discovered with $Planck$ through its bright submillimeter flux, originating from a pair of extraordinary dusty star-forming galaxies (DSFGs) at $z\approx 2.2$. Using JWST and interferometric mm/radio observations, we characterize the intrinsic physical properties of the DSFGs, which are separated by only $\sim 1^{\prime\prime}$ (8 kpc) and a velocity difference $ΔV \lesssim 600~{\rm km}~{\rm s}^{-1}$ in the source plane, and thus likely undergoing a major merger. Boasting intrinsic star formation rates ${\rm SFR}_{\rm IR} = 320 \pm 70$ and $400 \pm 80~ M_\odot~{\rm yr}^{-1}$, stellar masses ${\rm log}[M_\star/M_\odot] = 10.2 \pm 0.1$ and $10.3 \pm 0.1$, and dust attenuations $A_V = 1.5 \pm 0.3$ and $1.2 \pm 0.3$, they are remarkably similar objects. We perform spatially-resolved pixel-by-pixel SED fitting using rest-frame near-UV to near-IR imaging from JWST/NIRCam for both galaxies, resolving some stellar structures down to 100 pc scales. Based on their resolved specific SFRs and $UVJ$ colors, both DSFGs are experiencing significant galaxy-scale star formation events. If they are indeed interacting gravitationally, this strong starburst could be the hallmark of gas that has been disrupted by an initial close passage. In contrast, the host galaxy of the recently discovered triply-imaged SN H0pe has a much lower SFR than the DSFGs, and we present evidence for the onset of inside-out quenching and large column densities of dust even in regions of low specific SFR. Based on the intrinsic SFRs of the DSFGs inferred from UV through FIR SED modeling, this pair of objects alone is predicted to yield an observable $1.1 \pm 0.2~{\rm CCSNe~yr}^{-1}$, making this cluster field ripe for continued monitoring.

astro-ph.GA↗

arXiv2024

Almost Optically Dark Galaxies in DECaLS (I): Detection, Optical Properties and Possible Origins

We report the discovery of eight optical counterparts of ALFALFA extragalactic objects from DECaLS, five of which are discovered for the first time. These objects were flagged as HI emission sources with no optical counterparts in SDSS before. Multi-band data reveal their unusual physical properties. They are faint and blue ($g-r=-0.35\sim0.55$), with quite low surface brightness ($μ_{\rm g,peak}=24.88\sim26.41\,{\rm mag}/{\rm arcsec}^2$), irregular morphologies, low stellar masses ($log_{10}(M_{*}/M_\odot)=5.27\sim7.15$), low star formation rates ($SFR=0.21\sim9.24\times10^{-3}\,{M_\odot}\,{\rm yr}^{-1}$), and remarkably high HI-to-stellar mass ratios ($log_{10}(M_{\rm HI}/M_{*}) = 1.72\sim3.22$, except AGC\,215415). They deviate from the scaling relations between HI and optical properties defined by the ALFALFA sample and the baryonic Tully-Fisher relation. They agree well with the main sequence of star-forming galaxies but exhibit low star-forming efficiency. Based on their physical properties and environments, we speculate that six of these objects may have originated from tidal processes, while the remaining two appear to have isolated origins. They may have had a relatively calm evolutionary history and only begun to form stars recently.

astro-ph.GA↗

arXiv2024

Activating Wider Areas in Image Super-Resolution

The prevalence of convolution neural networks (CNNs) and vision transformers (ViTs) has markedly revolutionized the area of single-image super-resolution (SISR). To further boost the SR performances, several techniques, such as residual learning and attention mechanism, are introduced, which can be largely attributed to a wider range of activated area, that is, the input pixels that strongly influence the SR results. However, the possibility of further improving SR performance through another versatile vision backbone remains an unresolved challenge. To address this issue, in this paper, we unleash the representation potential of the modern state space model, i.e., Vision Mamba (Vim), in the context of SISR. Specifically, we present three recipes for better utilization of Vim-based models: 1) Integration into a MetaFormer-style block; 2) Pre-training on a larger and broader dataset; 3) Employing complementary attention mechanism, upon which we introduce the MMA. The resulting network MMA is capable of finding the most relevant and representative input pixels to reconstruct the corresponding high-resolution images. Comprehensive experimental analysis reveals that MMA not only achieves competitive or even superior performance compared to state-of-the-art SISR methods but also maintains relatively low memory and computational overheads (e.g., +0.5 dB PSNR elevation on Manga109 dataset with 19.8 M parameters at the scale of 2). Furthermore, MMA proves its versatility in lightweight SR applications. Through this work, we aim to illuminate the potential applications of state space models in the broader realm of image processing rather than SISR, encouraging further exploration in this innovative direction.

cs.CV↗

arXiv2024

EPOCHS IX. When cosmic dawn breaks: Evidence for evolved stellar populations in $7 < z < 12$ galaxies from PEARLS GTO and public NIRCam imaging

The presence of evolved stars in high-redshift galaxies can place valuable indirect constraints on the onset of star formation in the Universe. Thus we use PEARLS GTO and public NIRCam photometric data to search for Balmer-break candidate galaxies at $7 < z < 12$. We find that our Balmer-break candidates at $z \sim 10.5$ tend to be older (115 Myr), have lower inferred [O III] + H$β$ emission line equivalent widths (120 Å), have lower specific star formation rates (6 Gyr$^{-1}$) and redder UV slopes ($β= -1.8$) than our control sample of galaxies. However, these trends all become less strong at $z \sim 8$, where the F444W filter now probes the strong rest-frame optical emission lines, thus providing additional constraints on the current star formation activity of these galaxies. Indeed, the bursty nature of Epoch of Reionisation galaxies can lead to a disconnect between their current SED profiles and their more extended star-formation histories. We discuss how strong emission lines, the cumulative effect of weak emission lines, dusty continua and AGN can all contribute to the photometric excess seen in the rest-frame optical, thus mimicking the signature of a Balmer break. Additional medium-band imaging will thus be essential to more robustly identify Balmer-break galaxies. However, the Balmer break alone cannot serve as a definitive proxy for the stellar age of galaxies, being complexly dependent on the star-formation history. Ultimately, deep NIRSpec continuum spectroscopy and MIRI imaging will provide the strongest indirect constraints on the formation era of the first galaxies in the Universe, thereby revealing when cosmic dawn breaks.

astro-ph.GA↗

arXiv2024

EPOCHS Paper II: The Ultraviolet Luminosity Function from $7.5<z<13.5$ using 180 square arcminutes of deep, blank-fields from the PEARLS Survey and Public JWST data

We present an analysis of the ultraviolet luminosity function (UV LF) and star formation rate density of distant galaxies ($7.5 < z < 13.5$) in the `blank' fields of the Prime Extragalactic Areas for Reionization Science (PEARLS) survey combined with Early Release Science (ERS) data from the CEERS, GLASS, NGDEEP surveys/fields and the first data release of JADES. We use strict quality cuts on EAZY photometric redshifts to obtain a reliable selection and characterisation of high-redshift ($z>6.5$) galaxies from a consistently processed set of deep, near-infrared imaging. Within an area of 180 arcmin$^{2}$, we identify 1046 candidate galaxies at redshifts $z>6.5$ and we use this sample to study the ultraviolet luminosity function (UV LF) in four redshift bins between $7.5<z<13.5$. The measured number density of galaxies at $z=8$ and $z=9$ match those of past observations undertaken by the {\em Hubble Space Telescope} (HST). Our $z=10.5$ measurements lie between early JWST results and past HST results, indicating cosmic variance may be the cause of previous high density measurements. However, number densities of UV luminous galaxies at $z=12.5$ are high compared to predictions from simulations. When examining the star formation rate density of galaxies at this time period, our observations are still largely consistent with a constant star formation efficiency, are slightly lower than previous early estimations using JWST and support galaxy driven reionization at $z\leq8$.

astro-ph.GA↗

arXiv2024

How to coadd images: II. Anti-aliasing and PSF deconvolution

We have developed a novel method for co-adding multiple under-sampled images that combines the iteratively reweighted least squares and divide-and-conquer algorithms. Our approach not only allows for the anti-aliasing of the images but also enables PSF deconvolution, resulting in enhanced restoration of extended sources, the highest PSNR, and reduced ringing artefacts. To test our method, we conducted numerical simulations that replicated observation runs of the CSST/VST telescope and compared our results to those obtained using previous algorithms. The simulation showed that our method outperforms previous approaches in several ways, such as restoring the profile of extended sources and minimizing ringing artefacts. Additionally, because our method relies on the inherent advantages of least squares fitting, it is more versatile and does not depend on the local uniformity hypothesis for the PSF. However, the new method consumes much more computation than the other approaches.

astro-ph.IM↗

arXiv2024

100 Gbps Indoor Access and 4.8 Gbps Outdoor Point-to-Point LiFi Transmission Systems using Laser-based Light Sources

In this paper, we demonstrate the communication capabilities of light-fidelity (LiFi) systems based on highbrightness and high-bandwidth integrated laser-based sources in a surface mount device (SMD) packaging platform. The laserbased source is able to deliver 450 lumens of white light illumination and the resultant light brightness is over 1000 cd mm2. It is demonstrated that a wavelength division multiplexing (WDM) LiFi system with ten parallel channels is able to deliver over 100 Gbps data rate with the assistance of Volterra filter-based nonlinear equalisers. In addition, an aggregated transmission data rate of 4.8 Gbps has been achieved over a link distance of 500 m with the same type of SMD light source. This work demonstrates the scalability of LiFi systems that employ laserbased light sources, particularly in their capacity to enable highspeed short range, as well as long-range data transmission.

eess.SY↗

arXiv2024

Magellanic System Stars Identified in SMACS J0723.3-7327 James Webb Space Telescope Early Release Observations Images

We identify 71 distant stars in JWST/NIRCam ERO images of the field of galaxy cluster SMACS J0723.3-7327 (SMACS 0723). Given the relatively small ($\sim$$10^{\circ}$) angular separation between SMACS 0723 and the Large Magellanic Cloud, it is likely that these stars are associated with the LMC outskirts or Leading Arm. This is further bolstered by a spectral energy distribution analysis, which suggests an excess of stars at a physical distance of $40-100$ kpc, consistent with being associated with or located behind the Magellanic system. In particular, we find that the overall surface density of stars brighter than 27.0 mag in the field of SMACS 0723 is $\sim$2.3 times that of stars in a blank field with similar galactic latitude (the North Ecliptic Pole Time Domain Field), and that the density of stars in the SMACS 0723 field with SED-derived distances consistent with the Magellanic system is $\sim$6.1 times larger than that of the blank field. The candidate stars at these distances are consistent with a stellar population at the same distance modulus with [Fe/H] $= -1.0$ and an age of $\sim$$5.0$ Gyr. On the assumption that all of the 71 stars are associated with the LMC, then the stellar density of the LMC at the location of the SMACS 0723 field is $\sim$$740$ stars kpc$^{-3}$, which helps trace the density of stars in the LMC outskirts.

astro-ph.GA↗

arXiv2024

A search for high-redshift direct-collapse black hole candidates in the PEARLS north ecliptic pole field

Direct-collapse black holes (DCBHs) of mass $\sim 10^4$-$10^5 {M}_\odot$ that form in HI-cooling halos in the early Universe are promising progenitors of the $\gtrsim 10^9 {M}_\odot$ supermassive black holes that fuel observed $z \gtrsim 7$ quasars. Efficient accretion of the surrounding gas onto such DCBH seeds may render them sufficiently bright for detection with the JWST up to $z\approx 20$. Additionally, the very steep and red spectral slope predicted across the $\approx 1$-5 $μ$m wavelength range of the JWST/NIRSpec instrument during their initial growth phase should make them photometrically identifiable up to very high redshifts. In this work, we present a search for such DCBH candidates across the 34 arcmin$^{2}$ in the first two spokes of the JWST cycle-1 PEARLS survey of the north ecliptic pole time-domain field covering eight NIRCam filters down to a maximum depth of $\sim$ 29 AB mag. We identify two objects with spectral energy distributions consistent with the Pacucci et al. (2016) DCBH models. However, we also note that even with data in eight NIRCam filters, objects of this type remain degenerate with dusty galaxies and obscured active galactic nuclei over a wide range of redshifts. Follow-up spectroscopy would be required to pin down the nature of these objects. Based on our sample of DCBH candidates and assumptions on the typical duration of the DCBH steep-slope state, we set a conservative upper limit of $\lesssim 5\times 10^{-4}$ comoving Mpc$^{-3}$ (cMpc$^{-3}$) on the comoving density of host halos capable of hosting DCBHs with spectral energy distributions similar to the Pacucci et al. (2016) models at $z\approx 6$-14.

astro-ph.GA↗

arXiv2024

JWST's PEARLS: Improved Flux Calibration for NIRCam

The Prime Extragalactic Areas for Reionization and Lensing Science (PEARLS), a JWST GTO program, obtained a set of unique NIRCam observations that have enabled us to significantly improve the default photometric calibration across both NIRCam modules. The observations consisted of three epochs of 4-band (F150W, F200W, F356W, and F444W) NIRCam imaging in the Spitzer IRAC Dark Field (IDF). The three epochs were six months apart and spanned the full duration of Cycle 1. As the IDF is in the JWST continuous viewing zone, we were able to design the observations such that the two modules of NIRCam, modules A and B, were flipped by 180 degrees and completely overlapped each other's footprints in alternate epochs. We were therefore able to directly compare the photometry of the same objects observed with different modules and detectors, and we found significant photometric residuals up to ~ 0.05 mag in some detectors and filters, for the default version of the calibration files that we used (jwst_1039.pmap). Moreover, there are multiplicative gradients present in the data obtained in the two long-wavelength bands. The problem is less severe in the data reduced using the latest pmap (jwst_1130.pmap as of September 2023), but it is still present, and is non-negligible. We provide a recipe to correct for this systematic effect to bring the two modules onto a more consistent calibration, to a photometric precision better than ~ 0.02 mag.

astro-ph.IM↗

arXiv2024

Meta-Adapter: An Online Few-shot Learner for Vision-Language Model

The contrastive vision-language pre-training, known as CLIP, demonstrates remarkable potential in perceiving open-world visual concepts, enabling effective zero-shot image recognition. Nevertheless, few-shot learning methods based on CLIP typically require offline fine-tuning of the parameters on few-shot samples, resulting in longer inference time and the risk of over-fitting in certain domains. To tackle these challenges, we propose the Meta-Adapter, a lightweight residual-style adapter, to refine the CLIP features guided by the few-shot samples in an online manner. With a few training samples, our method can enable effective few-shot learning capabilities and generalize to unseen data or tasks without additional fine-tuning, achieving competitive performance and high efficiency. Without bells and whistles, our approach outperforms the state-of-the-art online few-shot learning method by an average of 3.6\% on eight image classification datasets with higher inference speed. Furthermore, our model is simple and flexible, serving as a plug-and-play module directly applicable to downstream tasks. Without further fine-tuning, Meta-Adapter obtains notable performance improvements in open-vocabulary object detection and segmentation tasks.

cs.CV↗

arXiv2024

Newly Formed Dust within the Circumstellar Environment of SNIa-CSM 2018evt

Dust associated with various stellar sources in galaxies at all cosmic epochs remains a controversial topic, particularly whether supernovae (SNe) play an important role in dust production. We report evidence of dust formation in the cold, dense shell behind the ejecta-circumstellar medium (CSM) interaction in the Type Ia-CSM SN 2018evt three years after the explosion, characterized by a rise in the mid-infrared (MIR) emission accompanied by an accelerated decline in the optical radiation of the SN. Such a dust-formation picture is also corroborated by the concurrent evolution of the profiles of the Ha emission line. Our model suggests enhanced CSM dust concentration at increasing distances from the SN as compared to what can be expected from the density profile of the mass loss from a steady stellar wind. By the time of the last MIR observations at day +1041, a total amount of 1.2+-0.2x10^{-2} Msun of new dust has been formed by SN 2018evt, making SN 2018evt one of the most prolific dust factories among SNe with evidence of dust formation. The unprecedented witness of the intense production procedure of dust may shed light on the perceptions of dust formation in cosmic history.

astro-ph.HE↗

arXiv2023

Individual Fairness under Uncertainty

Algorithmic fairness, the research field of making machine learning (ML) algorithms fair, is an established area in ML. As ML technologies expand their application domains, including ones with high societal impact, it becomes essential to take fairness into consideration during the building of ML systems. Yet, despite its wide range of socially sensitive applications, most work treats the issue of algorithmic bias as an intrinsic property of supervised learning, i.e., the class label is given as a precondition. Unlike prior studies in fairness, we propose an individual fairness measure and a corresponding algorithm that deal with the challenges of uncertainty arising from censorship in class labels, while enforcing similar individuals to be treated similarly from a ranking perspective, free of the Lipschitz condition in the conventional individual fairness definition. We argue that this perspective represents a more realistic model of fairness research for real-world application deployment and show how learning with such a relaxed precondition draws new insights that better explains algorithmic fairness. We conducted experiments on four real-world datasets to evaluate our proposed method compared to other fairness models, demonstrating its superiority in minimizing discrimination while maintaining predictive performance with uncertainty present.

cs.LG↗

arXiv2023

The Ly$α$ non-detection by JWST NIRSpec of a strong Ly$α$ emitter at $z=5.66$ confirmed by MUSE

The detections of Lyman-$α$ ($\rm Lyα$) emission in galaxies with redshifts above 5 are of utmost importance for constraining the cosmic reionization timeline, yet such detections are usually based on slit spectroscopy. Here we investigate the significant bias induced by slit placement on the estimate of $\rm Lyα$ escape fraction ( $f_{\rm esc}^{\mathrm{Lyα}}$), by presenting a galaxy (dubbed A2744-z6Lya) at $z=5.66$ where its deep JWST NIRSpec prism spectroscopy completely misses the strong $\rm Lyα$ emission detected in the MUSE data. A2744-z6Lya exhibits a pronounced UV continuum with an extremely steep spectral slope of $β=-2.574_{-0.008}^{+0.008}$, and it has a stellar mass of $\mathrm{\sim10^{8.82}~M_\odot}$, a star-formation rate of $\mathrm{\sim8.35~M_\odot yr^{-1}}$ and gas-phase metallicity of $\mathrm{12+log\,(O/H)\sim7.88}$. The observed flux and rest-frame equivalent width of its Ly$α$ from MUSE spectroscopy are $1.2\times \rm 10^{-16} erg~s^{-1}cm^{-2}$ and 75Å, equivalent to $f_{\rm esc}^{\mathrm{Lyα}}=78\pm4 \%$. However, its Ly$α$ non-detection from JWST NIRSpec gives a 5-$σ$ upper limit of $<13 \%$, in stark contrast to that derived from MUSE. To explore the reasons for this bias, we perform spatially resolved stellar population analysis of A2744-z6Lya using the JWST NIRCam imaging data to construct 2-dimensional maps of SFR, dust extinction and neutral hydrogen column density. We find that the absence of Ly$α$ in the slit regions probably stems from both the resonance scattering effect of neutral hydrogen and dust extinction. Through analyzing an extreme case in detail, this work highlights the important caveat of inferring $f_{\rm esc}^{\mathrm{Lyα}}$ from slit spectroscopy, particularly when using the JWST multiplexed NIRSpec microshutter assembly.

astro-ph.GA↗

arXiv2023

Criss Cross Nebula: Case study of shock regions with resolved microstructures at scales of $\sim$1000 AU

Using integral field spectroscopy from MaNGA, we study the resolved microstructures in a shocked region in Criss Cross Nebula (CCN), with an unprecedentedly high resolution of $\lesssim$1000 AU. We measure surface brightness maps for 34 emission lines, which can be broadly divided into three categories: (1) the [OIII] $λ$5007-like group including seven high-ionization lines and two [OII] auroral lines which uniformly present a remarkable lane structure, (2) the H$α$ $λ$6563-like group including 23 low-ionization or recombination lines which present a clump-like structure, and (3) [OII] $λ$3726 and [OII] $λ$3729 showing high densities at both the [OIII] $λ$5007 lane and the H$α$ clump. We use these measurements to constrain resolved shock models implemented in MAPPINGS V. We find our data can be reasonably well-fitted by a model which includes a plane-parallel shock with a velocity of $133\pm5$ km/s, plus an isotropic two-dimensional Gaussian component which is likely another clump of gas ionized by photons from the shocked region, and a constant background. We compare the electron density and temperature profiles as predicted by our model with those calculated using observed emission line ratios. We find different line ratios to provide inconsistent temperature maps, and the discrepancies can be attributed to observational effects caused by limited spatial resolution and projection of the shock geometry, as well as contamination of the additional Gaussian component. Implications on shock properties and perspectives on future IFS-based studies of CCN are discussed.

astro-ph.GA↗

arXiv2023

Star-forming Main Sequence of Giant Low Surface Brightness Galaxies

Giant Low Surface Brightness Galaxies (GLSBGs) are fundamentally distinct from normal galaxies (LSBGs) in star formation and evolution. In this work, we collected 27 local GLSBGs. They have high stellar masses (M*>10^10 Msolar) and low SFRs. With the specific SFRs lower than the characteristic value of the local star-forming (SF) galaxies of M*=10^10 Msolar(sSFR < 0.1 Gyr^-1), GLSBGs deviate from the SF main sequence (MS) defined for local SFGs respectively by E07 and S16 at the high M* regime. They are HI-rich systems with HI gas mass fractions (fHI) higher than the S16 MS galaxies, but have little molecular gas (H2), implying a low efficiency of HI-to-H2 transition due to the low HI surface densities that are far lower than the minimum of 6 - 8 Msolar pc^-2 required for shielding the formed H2 from photodissociation. For GLSBGs, the inner, bulge-dominated part with lower SFRs and higher M* is the main force pulling the entire GLSBG off from the MS, while the outer, disk-dominated part with relatively higher SFRs and lower M* reduces the deviations from the MS. For some cases, the outer, disk-dominated parts even tend to follow the MS. In the aspect of NUV - r versus g - r colors, the outer, disk-dominated parts are blue and behave similarly to the normal star-forming galaxies while the inner, bulge-dominated parts are in statistics red, indicating an inside-out star formation mechanism for the GLSBGs. They show few signs of external interactions in morphology, excluding the recent major merger scenario.

astro-ph.GA↗

arXiv2023

Skywork: A More Open Bilingual Foundation Model

In this technical report, we present Skywork-13B, a family of large language models (LLMs) trained on a corpus of over 3.2 trillion tokens drawn from both English and Chinese texts. This bilingual foundation model is the most extensively trained and openly published LLMs of comparable size to date. We introduce a two-stage training methodology using a segmented corpus, targeting general purpose training and then domain-specific enhancement training, respectively. We show that our model not only excels on popular benchmarks, but also achieves \emph{state of the art} performance in Chinese language modeling on diverse domains. Furthermore, we propose a novel leakage detection method, demonstrating that test data contamination is a pressing issue warranting further investigation by the LLM community. To spur future research, we release Skywork-13B along with checkpoints obtained during intermediate stages of the training process. We are also releasing part of our SkyPile corpus, a collection of over 150 billion tokens of web text, which is the largest high quality open Chinese pre-training corpus to date. We hope Skywork-13B and our open corpus will serve as a valuable open-source resource to democratize access to high-quality LLMs.

cs.CL↗