For example which question would result in more accurate responses: do you want lower taxes vs. do you want lower taxes if it cuts state budgets by 10%?
It's a bit leading, but not heavily. To be non-leading, it should also say that the high speed rail will take you from SF to LA in Y minutes, and tickets will cost $Z. Then it's non-leading; it's merely supplying additional information that not all poll respondents may know.
But it's not all that leading of a question. The cost and completion date, while they reflect badly on the project, are not irrelevant to peoples' opinion of it...
Do you think they would have got the same result if it was worded "A high speed train will reduce journey times by X%, reduce congestion by Y% on the roads and reduce CO2 pollution by Z%. Do you think we should build one?"
The full survey is here https://drive.google.com/file/d/1g5uibGxcEuknURkZZvT4Ah4Q9-I...