How is Javascript translated to bytecode?
It's up to the implementation; the specification is the full description of the language and how it's supposed to work, implementations are free to satisfy that implementation in any way they like. Some implementations seem (from the outside) to run it purely as an interpreter in the old sense; others may or may not compile to bytecode; V8 (the JavaScript engine in Chrome, Chromium, Brave, Node.js, and others) used to compile to machine code (twice, for hotspots in the app), but now starts out parsing to bytecode and running it in an interpreter and only compiling hotspots as necessary (details). (There's also a V8 mode where it only interprets, which they're experimenting with for environments where compiling at runtime isn't an option, such as iOS where non-Apple apps aren't allowed to allocate executable memory.)
The V8 team (V8 being the JavaScript engine in Chromium and Chrome) periodically publish descriptions of how they get the fantastic speed out of V8 that they do. You may find some of that on the V8 blog.
Naturally, you can also kick around the code of any of the open-source implementations. V8 and SpiderMonkey (Mozilla's engine) are the two major open-source ones I know.
This may help : http://www.ecma-international.org/publications/standards/Ecma-262.htm
There is no spec for how to translate into bytecode (That is up to the browser developers) but there are specs about how the language should behave
For Firefox there's some specifications on its bytecodes:
https://developer.mozilla.org/en-US/docs/Mozilla/Projects/SpiderMonkey/Bytecodeshttps://developer.mozilla.org/en-US/docs/Mozilla/Projects/SpiderMonkey/Internals/Bytecode
For V8 it's compiled to native code directly: