The D.C. Corrections Department said yesterday that it freed five prisoners it should not have under the city's controversial emergency prison release plan, but said that otherwise the program was implemented efficiently and effectively.

In his department's analysis of the early release program, D.C. Corrections Director Hallem H. Williams Jr. noted that only 58 of the 815 prisoners released between July 3 and Oct. 1 -- or 7.1 percent -- have been rearrested.

The prisoners released by mistake included one inmate, serving time for armed robbery, who was incorrectly given a 27-day reduction in his sentence, Williams said. The emergency law specifically prohibited the city from granting early release to prisoners convicted of armed robbery.

Two other inmates freed early under the program were serving mandatory minimum sentences -- a category specifically excluded from early release by the city's emergency law. And two more inmates were released earlier than they should have because of computation errors, according to the report.

But, other than those five mistakes, Williams maintained in the 20-page report released yesterday that "there is no clear evidence that early release of 815 inmates adversely affected public safety or increased the risk associated with release of any prisoners into the community."

The early release program, which allowed the city to free prisoners up to 90 days early, was implemented July 3 as a temporary measure to relieve the city's chronic prison crowding. The program ended Sept. 31, but a permanent early release program takes effect Nov. 17.

"I am generally pleased with the results of our analysis," Williams said. "It clearly indicates that Corrections Department personnel, working in cooperation with the Board of Parole, did an excellent job implementing the Emergency Powers Act."

Rep. Stan Parris (R-Va.), a longtime critic of the District's Lorton Reformatory in his congressional district in Fairfax County, attacked the report as being "totally wrong."

Parris said he randomly examined the records of 112 prisoners released under the program, and said his analysis showed that 80 percent of them were considered dangerous criminals under the D.C. criminal code.

He also claimed that 32 percent of them had their sentences illegally reduced because they were serving mandatory minimum sentences.

"We have all the persons identified, and all of their criminal records," said Parris. "The court records supplied to us by the Justice Department and the U.S. attorney's office simply do not compute with the facts that {Williams} uses in his report."

An attempt by Parris to overturn the early prisoner release program was defeated Oct. 15 by 10 votes in the Democratic-controlled House. Parris said that the closeness of the vote will force the city to tighten regulations for releasing prisoners.

Critics have contended that the District violated its commitment to release only prisoners convicted of nonviolent crimes.

In announcing the emergency program during the summer, Mayor Marion Barry said he wanted to reassure the public that "no prisoner convicted of a violent crime will have his sentence reduced under the emergency law."

Although robbery, attempted robbery and assault are considered violent crimes by the federal Bureau of Justice Statistics, and the D.C. Criminal Code includes robbery and assault in its definition of a "crime of violence," the Corrections Department has released prisoners convicted of robbery, attempted robbery, assault and weapons convictions. Yesterday's report did not deal with that discrepancy.

The report said that 50 percent of the prisoners released early were serving time for drug violations, and about 8 percent had been convicted of crimes against persons -- including robbery, assault, negligent homicide, cruelty to children and arson. Nearly 6 percent of the inmates had been convicted of weapons possession violations.

About 60 percent of the early release prisoners were serving time for misdemeanors.

Of the 815 prisoners released early, 88 percent, or 719 inmates, were serving their full sentences and did not have to appear before the D.C. Parole Board, the report said. The inmates were released an average of 19 days early; nearly 90 percent of them would have been released by Oct. 14 without the plan, according to the report.