2 MAINAK SINGHA ET AL. violet (UV) sources play a key role in this process (Robertson et al. 2015; Naidu et al. 2020; Yung et al. 2020a,b). Nev- ertheless, it is imperative that AGN can only dominate the reionization energy budget if there is an AGN abundance at z> 5. Recent JWST observations have reported an overdensity of AGN within the redshift range z =4 − 12. These AGN span a broad range in their Eddington ratios (λ Edd =0.06 − 2), bolometric luminosities (L bol = 10 43 − 10 46 erg s −1 ), and black hole masses (M BH = 10 5 − 10 8 M ⊙ )(Harikane et al. 2023; Larson et al. 2023; Juodˇ zbalis et al. 2023). Studies em- ploying JWST/NIRSpec observations propose a lower limit for unobscured AGN fractions between z =6.5 − 12, esti- mated at approximately 5% (Juodˇ zbalis et al. 2023; Xu et al. 2023). A recent JWST/ MIRI imaging study reported that around 25% of galaxies in their sample at redshifts z =3 − 5 harbor heavily obscured AGN and composite AGN (Yang et al. 2023). This discovery suggests a threefold increase in the rate of black hole growth at these redshifts, exceeding expectations based on X-ray data. Such findings have sub- stantial implications for our understanding of cosmic reion- ization, as accelerated black hole growth may imply an over- density at these redshifts. Additionally, popular black hole growth models suggest that unobscured AGN is a direct re- sult of the obscuring gas clouds being blown out by the AGN- driven superwinds (Alexander & Hickox 2012). However, we need to keep in mind that at z> 5: (1) AGN observations are relatively scarce, and (2) dwarf starburst galaxies dom- inate the overall galaxy population (Robertson et al. 2015; Atek et al. 2024). Therefore, local analogs of z> 5 AGN host galaxies provide us with an attractive opportunity to bet- ter predict the presence of the first black holes, understand their properties and growth scenarios, and quantify their po- tential role in cosmic reionization. Nearby (z ∼ 0.3), low-metallicity dwarf galaxies also known as green peas (Cardamone et al. 2009) have garnered significant attention due to their remarkable similarities with z> 5 Lyα galaxies as seen from recent JWST spectroscopic observations (Rhoads et al. 2023). Such similarities include having subsolar metallicities, [O III]5007 ˚ A rest-frame equiv- alent width exceeding 500 ˚ A half-light radius, r 50 < 0.3 kpc, and an interacting nature (Izotov et al. 2011b; Malhotra et al. 2012; Henry et al. 2015; Yang et al. 2016). In the past two decades, X-ray observations have exten- sively searched for AGN in Lymanα emitters at redshifts z> 2 (Malhotra et al. 2003; Basu-Zych & Scharf 2004; Wang et al. 2004; Yang et al. 2009; Zheng et al. 2012; Calhau et al. 2020). A commonly used indicator for AGN identification is the 2-10 keV X-ray luminosity, L 2−10 keV > 10 42 erg s −1 (Nandra et al. 2002; Haines et al. 2012; Birchall et al. 2022). However, solely focusing on high-luminosity AGN might overlook those with lower luminosities (L 2−10 keV < 10 42 erg s −1 ). For instance, our Milky Way’s supermassive black hole, Sgr A*, has a quiet-state X-ray luminosity of L 2−10 keV ∼ 10 33−35 erg s −1 (Sabha et al. 2010). Recent IXPE observations have shown flaring events boosting its X-ray luminosity to L 1−100 keV ∼ 10 39−44 erg s −1 (Marin et al. 2023). Other studies have identified LLAGN with L 2−10 keV ∼ 10 37−42 erg s −1 (She et al. 2018; Diaz et al. 2020). Additionally, X-ray coronal destruction events can drastically reduce X-ray luminosity, as seen in a changing- look AGN (Ricci et al. 2020). However, lowering the luminosity threshold may misiden- tify non-AGN sources such as super-Eddington accretors, which are essentially stellar mass black holes or neutron stars and are a class of high-mass X-ray binaries (HMXBs). In low-metallicity dwarf galaxies, HMXBs contribute signifi- cantly, with luminosities reaching 10 41 ergs −1 (Lehmer et al. 2021, 2022). However, these HMXBs could be either super- Eddington accretors or IMBH/LLAGN, with their 2-10 keV luminosities often exceeding 10 39 ergs −1 (Swartz et al. 2011; Kaaret et al. 2017), which are often termed ultraluminous X-ray sources (ULXs). Despite being long believed to host IMBHs, there has been a plethora of evidence showing ULXs may actually contain super-Eddington neutron star accretors (Sutton et al. 2013; Bachetti et al. 2014; Karino & Miller 2016; Israel et al. 2017; Pintore et al. 2018). Therefore, re- lying solely on X-ray observations cannot definitively deter- mine the presence of LLAGN, necessitating additional con- straints. Optical emission lines such as Hα, and He IIλ4686, com- bined with X-ray observations could be an incredibly power- ful tool to constrain the origin of X-ray emission (Fabrika et al. 2015; Lin et al. 2018). For example, Fabrika et al. (2015) found that the line-widths of He IIλ4686 and Hα shows almost a one-to-one correlation for super-Eddington accretors, where such high accretion rate almost always launches a very strong wind (Shakura & Sunyaev 1973). In this paper, we adopt this approach to investigate whether green peas truly host low luminosity AGN. This paper is organized as follows: In Section 2, we first briefly describe the data. We then investigate the multi- wavelength data and outline the main results in Section 3, while in Section 4 we combine the results from the observa- tions and attempt to explain the origin of any observed multi- wavelength emission mechanisms and their potential connec- tion to AGN. Finally in Section 5 we present our conclusions. 2. OBSERVATION AND DATA 2.1. Sample Selection: We began with the parent sample of green pea galaxies from Jiang et al. (2019). This consists of 1004 objects se- lected from Data Release 13 of the Sloan Digital Sky Survey (SDSS DR13). The primary selection was for high equiva-