Your post is unclear and confusing. In some places you assume a marginally employable worker that is employed. Elsewhere you assume a not employed person.
You seem to be asking if a not employed person receiving UBI creates value for society. I will answer to that in the hope that it is a succinct re-expression of your intended question. I will ignore non-market production such as what occurs in a household where for example a spouse produces valuable child care and remains at home while the other spouse produces value by working at a company.
If we cannot determine that a person consistently destroys value in an employment situation then we will continually value the avoidance of value destruction. Since generally we can make that determination, because we can wisen-up, this kind of accounting seems unnecessary beyond a short term mistake. I think your question is not focused on the short term. I think a longer term focus suggests we will not value the avoidance of value destruction. One way to understand this is to determine the price an employer is willing to pay to avoid the loss created by the value destroying employee. It is initially as much as the value that could be destroyed but after deciding to leave that person idle, the price an employer is willing to pay becomes zero. The decision is sufficient to solve the problem.
I think it makes sense to assume that an idle person is not creating value, rather then to try to account for costs and benefits of employment situations that did not happen. This serves as my answer to your question that asks "if an unemployed person actually provides more benefit to society by not working than working". Edit: In your comment you posit that a young Einstein entertains people by playing video games and makes science discoveries. Entertainers are not idle. Doing science is not being idle. This part of my answer concerns people that are strictly idle.
You inquired whether individual consumption is valuable to society. By definition consumption cannot create a benefit for society apart from the benefit to the consumer unless you imagine the individual is consuming something bad like a toxic waste that is an environmental hazard. Usually we assume the individual is consuming something good, like an apple, and once it is consumed, that apple cannot be consumed again. In that one act of consumption, involving one individual and one apple, society did not enter the picture so there no benefit to society apart from the benefit to the consumer. Individual consumption is the enjoyment of value regardless of whether the person is idle or producing.
For an individual, lifetime consumption and lifetime production do not have to match because governments make transfers. Since they don't have to match, you cannot even use consumption as some sort of estimate of the value that they might create in production. That answers your question "can an individual's value to society be defined by their consumption"?.
You inquired whether consumption drives economic health. When economists are concerned about a lack of consumption, lack of investment, or lack of demand, they are usually commenting about a short term mismatch between the qualities and quantities of what is supplied and the qualities and quantities of what is demanded, such as in a recession. I think your question is not focused on the short term. Since it is not short term, I suggest we would not be focused on whether adjusting consumption drives anything at all. In the long term a high level of consumption is a "picture" of economic health.
You wrote... "I'm looking for scientific studies where humans working in a "dirty job" has been shown to provide some intrinsic benefit over that "dirty job" being automated." I am uncertain what you mean but I interpreted this inquiry to be asking whether there is economic justification for or against the automation of dirty jobs. My comment is that whether it is economical to automate is particular to the situation; there is no universal answer and the nature of the analysis involves costs and benefits. Dirtiness can impose costs. The fact of dirtiness is subsumed in the cost consideration so from the abstract perspective, the analysis of dirty jobs is not different from other analysis.
Your post makes what seems to me to be multiple distinct inquiries. This forum discourages multiple questions in a single post.