For better or worse, Jack Gillis, late of the National Highway Traffic Safety Administration, has added his name to the roll of born-again bureaucrats who defied their agencies to give the public information they felt it wouldn't otherwise get.

In Gillis' case, the information was "The Car Book," a compilation of government data on the fuel efficiency, crash-worthiness and maintenance cost of cars sold in the United States. Last year, Gillis, a $40,000-a-year GS14, was part of a NHTSA team that put out the 1981 car book. After the Reagan administration announced it would discontinue the book, Gillis got together with a private publisher and decided to do the car book for 1982--in secret and on his own time.

It all has heroic overtones--truth vs. cover-up, the individual vs. the institution. A personal attack by NHTSA's current administrator, Raymond Peck, didn't hurt. The venture's commercial aspect--the new book sells for $4.95--does little to diminish the story's appeal. Within days, Gillis had resigned his job as a marketing specialist and become a media personality, appearing on the Phil Donahue Show (to be aired here tomorrow), the Today show, and assorted radio and television broadcasts in New York and Washington.

But hidden amidst the hoopla is one troubling question, the question of competing allegiances that career government employes must deal with as administrations come and go. To whom do they owe their loyalty? Political bosses? Individual colleagues or subordinates? Their agency or its mission? The public? Themselves? If the answer is some or all of the above, how is that loyalty divided?

Gillis sees it this way: "In a government bureaucracy, loyalty is to the people, the taxpayer, not to a political appointee or supervisor. This information belongs to the public."

But the people Gillis left behind at NHTSA see it differently. They say they were also trying to get the information out to the public--maybe not in the same form as "The Car Book"--but in other ways. Gillis was part of this effort, and he says he still believes in it. But he believed in the car book more.

"I was disappointed in Gillis ," said Mike Finkelstein, the NHTSA associate administrator over Gillis. "Jack Gillis is a talented person . . . . Talented people working together over the long haul can change institutions. As a person trying to make an institution run properly, it's upsetting when people you count on to help you do that take a walk . . . .

"If I agreed there was some kind of cover-up . . . it would be much easier" to understand what Gillis did, said Finkelstein, who has worked at the Transportation Department for 13 years, five of them at NHTSA.

But, he added, the decision whether to publish "The Car Book" was "like a lot of calls--subjective." The format and validity of the book had been debated before Ronald Reagan came to office: it was questioned whether one-time head-on crash tests provided accurate data about the dangers to a car's occupants. True, the auto industry was no fan of the book and may have helped kill it. But NHTSA officials were working to get the information out in other ways, such as the agency's hotline.

"The problem I have is with the arrogance," Finkelstein continued. "Where some people impose their morality and say this issue is black and white, there is a wrong and a right . . . . Where I get mad is when he says it's a black and white decision . . . because he knows it's not."

Gillis, at 31 a three-year veteran of NHTSA (earlier he worked as an aide to Western Union's vice president for consumer relations), does regard the failure to reissue the car book as a cover-up. But he was surprised at the intensity of the department's reaction, the vituperation, the news conference at which Peck said Gillis' data was unreliable and old, the move to have DOT's inspector general look into the case.

Why was Peck upset? Simply at the public embarrassment of one of his employes flouting his decision? "I really don't know," said Gillis. "He always suggested this the book should be done on a private basis. That's what I did.

"I think they felt I had an obligation to tell them I was doing this. I don't have an obligation to tell them anything that I'm doing on my private time."

Why keep it a secret? In part because he doubted he could pull the project off, in part on the advice of his publisher, Tilden Press' Joel Makower, who knew other companies were interested in publishing such a book and all the information was publicly available.

"In hindsight, I'm sure glad I didn't tell anyone because who knows what they would have done," he said.

"I'm bothered by the secrecy, the distrust of us," said Mike Brownlee, a NHTSA office director who was Gillis's boss once removed. "And I'm bothered by the question of his influencing decisions around here that may have a bearing on his own personal gain. I don't think there would be any problem if he had quit when he began work on the book , or if he had informed people he was going to do a private project."

"I don't believe you're never allowed to speak out," added Finkelstein. "I do believe you're not allowed to when you're taking their money."

Other NHTSA employes are not so critical --perhaps a dozen have phoned him in support. Even Brownlee and Finkelstein have wished him well in his new work: Gillis has formed a one-man consulting firm "to sell the profitability of consumerism to the corporate sector."

He says he has no second thoughts. "I'm very glad I did it. There is no departmental or legal rule that I violated . . . . I would much rather have done this as a government employe. I'd love to see NHTSA do a 1983 version. But," he said, "I don't think they will."