I suppose one way to assess whether or not knowledge has an inherent value all of its own is to see whether or not it applies to all knowledge. We could, for example, have a thread on this forum in which each poster posts once a day, describing in great detail (no fewer than 1,000 words per item) what's on the top shelf of their fridge. It could even be in the forum rules that everybody must make this post first thing and include a picture as proof otherwise they are not allowed to make any other posts, and failure to comply with this rule is an instant perma-ban.
Assuming there wasn't a mass exodus with people forming an off-shoot sister site without such a rule, this would quickly become a large repository of knowledge. But not a repository that has any scientific value - it's a self-selecting group, it's self-reporting so people can easily lie, and even if those things weren't true the data as a whole would have little even sociological value as it is, at best, incomplete (both in terms of what's in the fridge as a whole, and in terms of demographical information) and isn't attached to any particular hypothesis. It also seems that it would have limited value as a source of interest - some people may be vaguely interested at first, but I think the novelty would wear off quite quickly, and I think that the number of people who would maintain an interest would be very low.
So does this knowledge have inherent value, simply because it's knowledge? If not, how do we determine what knowledge has inherent value, and which doesn't?