Here is the story about the genesis of the IBM 701.

The IBM 701 arrived at Livermore about 1954. Hardware floating point, index registers, the stack, linked lists were all in the future! Here are a few comments about it:
Short note on early use
Doing without floating point
Brief comment in passing on Livermore’s early compiler
Note on Calling Conventions
Music for the 701
Doing Input Output on early IBM computers
Summing an Array
Fixing the 701
Scheduling memory refreshes for the Williams Tubes

The standard memory was Williams tubes which provided 2K words of 36 bits each. Many customers doubled that to 4K which required what might have been the first “small address hack”.

Notes by others

The Principles of Operation! More essential 701 material
Much detailed information on the machine including reliability and cost of operation.
A picture and some notes
701 at Livermore
IBM provides some good and interesting history.
Gordon Bell provides a perspective on the broader series of 36 bit machines from IBM.
Some more history
The 704 succeeded the 701 in IBM’s 36 bit series.
Dura A. Sweeney, “Los Alamos Coding System and Assembly Program for the IBM 701,” IEEE Annals of the History of Computing, vol. 05, no. 2, pp. 178-180, Apr-Jun, 1983.
Edward A. Voorhees, “Los Alamos Debugging Programs and Techniques as Used on the IBM 701,” IEEE Annals of the History of Computing, vol. 05, no. 2, pp. 180-182, Apr-Jun, 1983.
Inventors of the Modern Computer
Dual was an primitive assembler developed at Los Alamos and used at Livermore.
More detail on instructions