While most scientists at this federal laboratory are designing new nuclear weapons or inspecting old ones, a few scientists busily work on technology that might someday help put their colleagues out of business.

The Lawrence Livermore National Laboratory is one of the government's two nuclear weapon design labs, but it is also a center of expertise on arms control verification. Scientists here look for ways to detect and measure nuclear explosions, so that they could detect Soviet cheating if the two superpowers ever agreed to stop testing atomic weapons.

The two nations signed a Limited Test Ban Treaty in 1963 that allows underground tests but prohibits blasts in space, underwater or in the atmosphere. Soviet leader Mikhail Gorbachev proposed a comprehensive test ban treaty last year, but the Reagan administration has demurred.

Thanks to new and more precise ways of measuring the shock waves that bomb blasts send through the earth, Livermore scientists believe that they could detect almost any explosion if they were allowed to station listening posts in Soviet territory.

"You can define a verification scheme that can give you confidence that there aren't any tests above one kiloton," Milo D. Nordyke, chief of the lab's verification program, said in a recent interview.

Tests of weapons less powerful than one kiloton wouldn't be all that useful to the Soviet military, experts here said. Still, they caution that it will be difficult to absolutely preclude cheating.

Paul S. Brown, the laboratory's arms control chief who opposes a comprehensive test ban, suggested that the Soviets might violate a treaty by testing nuclear weapons deep in outer space.

"They could certainly get away with it," he said. "They could go beyond Mars, in which case we'd have to go beyond Mars to measure it."

But most of the research here focuses on detecting underground tests, and even test ban opponents acknowledge that substantial strides have been made in such seismic monitoring. The lab helped design an array of detectors in Norway that can measure explosions in the Soviet Union with unprecedented accuracy, Nordyke said.

To police a test ban, the United States would need to deploy many such devices, designed to cope with a number of potential problems. One is the Soviet program of "peaceful" nuclear explosions.

During the past two decades, the Soviets have conducted more than 70 blasts to stimulate gas production, for geological exploration in Siberia or to create huge underground storage cavities. The United States abandoned a more limited program to look for peaceful uses of atomic explosions in the mid-1970s.

U.S. officials worry that the Soviets might gain some militarily useful knowledge from such explosions if a weapons test ban were in effect. At the same time, the Soviets might use a large peaceful explosion to conceal the simultaneous detonation of a smaller-yield bomb, officials say.

Similarly, detection devices would have to be able to distinguish between earthquakes and atomic blasts and to ensure that the Soviets did not take advantage of a large earthquake to detonate an atomic weapon.

Nordyke also said that the Soviets could "decouple," or muffle, an atomic explosion in a huge underground cavity whose walls would absorb most of the shock waves. A 10-kiloton blast in such a cavity might register as less than a one-kiloton blast on U.S. monitors.

Experts here believe that modern technology can solve most of these problems, but not with absolute certainty; politicians will have to decide the range of uncertainty they would accept. Thus, Nordyke's colleague W. J. Hannon wrote in Science magazine that a system of 15 seismic arrays inside the Soviet Union could detect even three- to 10-kiloton blasts in decoupled cavities with 90 percent confidence.

During negotiations between 1978 and 1980, the Soviets agreed in principle to allow detectors to be placed in the Soviet Union. But Nordyke said that the search for the best sites would pose delicate political and technical problems.

Nordyke added, however, that he is more concerned that the United States has devoted little effort to developing technology to monitor more traditional arms control treaties that would limit production or deployment of nuclear missiles.

Although "verification" and "on-site inspection" have been buzzwords of the Reagan administration's arms control policy, Nordyke said, little thought or research has been devoted to how such policies would be carried out.

"What does that mean? What equipment would you take? What would you look for?" he asked. "There's been very little work that's gone beyond the phrase . . . .It isn't a matter of just going over there and kicking rocks."