An older example I'd posted:

http://www.ethicalhacker.net/component/option,com_smf/Itemid,54/topic,7988.msg42741/#msg42741

Hope it helps.

@hayabusa
Thanks. I highly appreciate it,
The thing is that I need not one or two but a few thousand samples found in the wild. Generating them individually would consume a lot of time. That's why I was asking if someone already has a database maybe they could share it.

If I'm wanting to quickly analyse some JS in the wild I usually turn to Wepawet. I've no affiliation with the service but it couldn't hurt to get in touch with the team there to see if they're willing/able to provide access to some of their samples?

Alternatively, some of Wepawet's reports can be accessed based on md5 hash of the content (I found this report via a quick google search for example). I've not read their Ts&Cs so use at your own risk, but a quick Google Dork of:
is currently returning >15k results

Unfortunately Wepawet's report format only lists the de-obfuscated operations rather than the original source so may not be exactly relevant to your needs, but you could always use the listed report targets to grab any scripts that are still live yourself.

Hope this helps, good luck with your project.

Not sure if you can see recent submission on iseclab's wepawet site. Here's another one work a look, use the search feature to grab more recently checked URLs

http://urlquery.net/search.php?q=.&type=string&start=2012-09-24&end=2012-09-26&max=50

You might need to sift through some of the lower repped results to get some obsfucated javascript. And they may still be up.

Congrats, enjoy your studies!