{"id":3757,"date":"2022-03-15T19:40:29","date_gmt":"2022-03-15T19:40:29","guid":{"rendered":"https:\/\/billsbud.org\/?p=3757"},"modified":"2022-03-15T19:40:29","modified_gmt":"2022-03-15T19:40:29","slug":"new-gpt-3-capabilities-edit-insert","status":"publish","type":"post","link":"https:\/\/billsbud.org\/?p=3757","title":{"rendered":"New GPT-3 Capabilities: Edit & Insert"},"content":{"rendered":"

<\/p>\n

\n\"New<\/p>\n

We’ve released new versions of GPT-3<\/a> and Codex<\/a> which can edit or insert content into existing text, rather than just completing existing text. These new capabilities make it practical to use the OpenAI API<\/a> to revise existing content, such as rewriting a paragraph of text or refactoring code. This unlocks new use cases and improves existing ones; for example, insertion is already being piloted in GitHub Copilot<\/a> with promising early results.<\/p>\n<\/div>\n

\n Read Edit Documentation<\/a><\/p>\n
\n Read Insert Documentation<\/a><\/p>\n
\n Try in Playground<\/a>
\n<\/section>\n
\n
\n
\n
\n
\n
\n
 <\/div>\n
def___<\/span>\n
fib(10)<\/code><\/pre>\n<\/div>\n
\n
 <\/div>\n
def fib(n):\n    if n <= 1 1: return fib(n-1) + fib(n-2) < span>
fib(10)<\/=><\/span><\/code><\/pre>\n<\/div>\n
\n
 <\/div>\n
def fib(n):\n    if n <= 1 1: return fib(n-1) + fib(n-2) < span>
fib(10)<\/=><\/span><\/code><\/pre>\n<\/div>\n
\n
Improve<\/div>\n
def fib(n):\n    if n <= 1 1: return fib(n-1) + fib(n-2) < span>
fib(10)<\/=><\/span><\/code><\/pre>\n<\/div>\n
\n
Improve the<\/div>\n
def fib(n):\n    if n <= 1 1: return fib(n-1) + fib(n-2) < span>
fib(10)<\/=><\/span><\/code><\/pre>\n<\/div>\n
\n
Improve the runtime<\/div>\n
def fib(n):\n    if n <= 1 1: return fib(n-1) + fib(n-2) < span>
fib(10)<\/=><\/span><\/code><\/pre>\n<\/div>\n
\n
Improve the runtime complexity<\/div>\n
def fib(n):\n    if n <= 1 1: return fib(n-1) + fib(n-2) < span>
fib(10)<\/=><\/span><\/code><\/pre>\n<\/div>\n
\n
Improve the runtime complexity of the<\/div>\n
def fib(n):\n    if n <= 1 1: return fib(n-1) + fib(n-2) < span>
fib(10)<\/=><\/span><\/code><\/pre>\n<\/div>\n
\n
Improve the runtime complexity of the function<\/div>\n
def fib(n):\n    if n <= 1 1: return fib(n-1) + fib(n-2) < span>
fib(10)<\/=><\/span><\/code><\/pre>\n<\/div>\n
\n
Improve the runtime complexity of the function<\/div>\n
def fib(n, memo={}):\n    if n in memo:\n        return memo[n]\n    if n <= 1: f=\"1\" else: memo) + fib(n-2, memo[n]=\"f\" return f< code><\/=><\/code><\/pre>\n<\/div>\n
\n
 <\/div>\n
def fib(n, memo={}):\n    if n in memo:\n        return memo[n]\n    if n <= 1: f=\"1\" else: memo) + fib(n-2, memo[n]=\"f\" return f< span><\/=><\/span><\/code><\/pre>\n<\/div>\n
\n
Translate<\/div>\n
def fib(n, memo={}):\n    if n in memo:\n        return memo[n]\n    if n <= 1: f=\"1\" else: memo) + fib(n-2, memo[n]=\"f\" return f< span><\/=><\/span><\/code><\/pre>\n<\/div>\n
\n
Translate to<\/div>\n
def fib(n, memo={}):\n    if n in memo:\n        return memo[n]\n    if n <= 1: f=\"1\" else: memo) + fib(n-2, memo[n]=\"f\" return f< span><\/=><\/span><\/code><\/pre>\n<\/div>\n
\n
Translate to JavaScript<\/div>\n
def fib(n, memo={}):\n    if n in memo:\n        return memo[n]\n    if n <= 1: f=\"1\" else: memo) + fib(n-2, memo[n]=\"f\" return f< span><\/=><\/span><\/code><\/pre>\n<\/div>\n
\n
Translate to JavaScript<\/div>\n
function fibonacci(n) {\n    var memo = {};\n    return (function fib(n, memo) {\n        if (n in memo) return memo[n];\n        var f;\n        if (n <= 1) f=\"1;\" else memo) + fib(n-2, memo); memo[n]=\"f;\" return f; })(n, }< code><\/=><\/code><\/pre>\n<\/div>\n
\n
 <\/div>\n
function fibonacci(n) {\n    var memo = {};\n    return (function fib(n, memo) {\n        if (n in memo) return memo[n];\n        var f;\n        if (n <= 1) f=\"1;\" else memo) + fib(n-2, memo); memo[n]=\"f;\" return f; })(n, }< code><\/=><\/span><\/code><\/pre>\n<\/div>\n
\n
Write<\/div>\n
function fibonacci(n) {\n    var memo = {};\n    return (function fib(n, memo) {\n        if (n in memo) return memo[n];\n        var f;\n        if (n <= 1) f=\"1;\" else memo) + fib(n-2, memo); memo[n]=\"f;\" return f; })(n, }< code><\/=><\/span><\/code><\/pre>\n<\/div>\n
\n
Write this<\/div>\n
function fibonacci(n) {\n    var memo = {};\n    return (function fib(n, memo) {\n        if (n in memo) return memo[n];\n        var f;\n        if (n <= 1) f=\"1;\" else memo) + fib(n-2, memo); memo[n]=\"f;\" return f; })(n, }< code><\/=><\/span><\/code><\/pre>\n<\/div>\n
\n
Write this as<\/div>\n
function fibonacci(n) {\n    var memo = {};\n    return (function fib(n, memo) {\n        if (n in memo) return memo[n];\n        var f;\n        if (n <= 1) f=\"1;\" else memo) + fib(n-2, memo); memo[n]=\"f;\" return f; })(n, }< code><\/=><\/span><\/code><\/pre>\n<\/div>\n
\n
Write this as a<\/div>\n
function fibonacci(n) {\n    var memo = {};\n    return (function fib(n, memo) {\n        if (n in memo) return memo[n];\n        var f;\n        if (n <= 1) f=\"1;\" else memo) + fib(n-2, memo); memo[n]=\"f;\" return f; })(n, }< code><\/=><\/span><\/code><\/pre>\n<\/div>\n
\n
Write this as a one<\/div>\n
function fibonacci(n) {\n    var memo = {};\n    return (function fib(n, memo) {\n        if (n in memo) return memo[n];\n        var f;\n        if (n <= 1) f=\"1;\" else memo) + fib(n-2, memo); memo[n]=\"f;\" return f; })(n, }< code><\/=><\/span><\/code><\/pre>\n<\/div>\n
\n
Write this as a one-liner<\/div>\n
function fibonacci(n) {\n    var memo = {};\n    return (function fib(n, memo) {\n        if (n in memo) return memo[n];\n        var f;\n        if (n <= 1) f=\"1;\" else memo) + fib(n-2, memo); memo[n]=\"f;\" return f; })(n, }< code><\/=><\/span><\/code><\/pre>\n<\/div>\n
\n
Write this as a one-liner<\/div>\n
function fibonacci(n) {\n    var memo = {};\n    return (function fib(n, memo) {\n        return n in memo ? memo[n] : (memo[n] = n <= 1 ? : fib(n-1, memo) + fib(n-2, memo)); })(n, memo); }< code><\/=><\/code><\/pre>\n<\/div>\n
\n
 <\/div>\n

function fibonacci(n) {\n    var memo = {};\n    return (function fib(n, memo) {\n        return n in memo ? memo[n] : (memo[n] = n <= 1 ? : fib(n-1, memo) + fib(n-2, memo)); })(n, memo); }< code><\/=><\/code><\/pre>\n<\/div>\n
\n
 <\/div>\n
\/\nfunction fibonacci(n) {\n    var memo = {};\n    return (function fib(n, memo) {\n        return n in memo ? memo[n] : (memo[n] = n <= 1 ? : fib(n-1, memo) + fib(n-2, memo)); })(n, memo); }< code><\/=><\/code><\/pre>\n<\/div>\n
\n
 <\/div>\n
\/**\nfunction fibonacci(n) {\n    var memo = {};\n    return (function fib(n, memo) {\n        return n in memo ? memo[n] : (memo[n] = n <= 1 ? : fib(n-1, memo) + fib(n-2, memo)); })(n, memo); }< code><\/=><\/code><\/pre>\n<\/div>\n
\n
 <\/div>\n
\/**___<\/span>\nfunction fibonacci(n) {\n    var memo = {};\n    return (function fib(n, memo) {\n        return n in memo ? memo[n] : (memo[n] = n <= 1 ? : fib(n-1, memo) + fib(n-2, memo)); })(n, memo); }< code><\/=><\/code><\/pre>\n<\/div>\n
\n
 <\/div>\n
\/**​\n * Recursive Fibonacci function with memoization.\n * @param {number} n\n * @returns {number}\n *\/<\/span>\nfunction fibonacci(n) {\n    var memo = {};\n    return (function fib(n, memo) {\n        return n in memo ? memo[n] : (memo[n] = n <= 1 ? : fib(n-1, memo) + fib(n-2, memo)); })(n, memo); }< code><\/=><\/code><\/pre>\n<\/div>\n<\/div>\n
\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n
GPT-3 and Codex have traditionally added text to the end of existing content, based on the text that came before. Whether working with text or code, writing is more than just appending—it’s an iterative process where existing text is revised. GPT-3 and Codex can now edit text, changing what’s currently there or adding text to the middle of content.<\/p>\n
Inserting into the middle of text<\/h2>\n
The new insert<\/a> capability<\/a> adds contextually relevant text in the middle of existing content. Providing future context to the model can improve the quality of completions for applications such as writing long-form text, transitioning between paragraphs, following an outline, or guiding the model toward an ending.<\/p>\n
\n
\n
\n
\n
\n
\n
\nSection 1. High school graduation<\/p>\n
Today's the big day!___<\/span>\n<\/div>\n<\/div>\n
\n
\nSection 1. High school graduation<\/p>\n
Today's the big day! Congratulations to all our high school graduates!<\/p>\n
We hope you enjoy your graduation ceremonies and wish you the best of luck in your future endeavors.<\/span>\n<\/div>\n<\/div>\n
\n
\nSection 1. High school graduation<\/p>\n
Today's the big day! Congratulations to all our high school graduates!<\/p>\n
We hope you enjoy your graduation ceremonies and wish you the best of luck in your future endeavors.<\/span><\/p>\n
Section\n<\/p><\/div>\n<\/div>\n
\n
\nSection 1. High school graduation<\/p>\n
Today's the big day! Congratulations to all our high school graduates!<\/p>\n
We hope you enjoy your graduation ceremonies and wish you the best of luck in your future endeavors.<\/span><\/p>\n
Section 2.\n<\/p><\/div>\n<\/div>\n
\n
\nSection 1. High school graduation<\/p>\n
Today's the big day! Congratulations to all our high school graduates!<\/p>\n
We hope you enjoy your graduation ceremonies and wish you the best of luck in your future endeavors.<\/span><\/p>\n
Section 2. Moving\n<\/p><\/div>\n<\/div>\n
\n
\nSection 1. High school graduation<\/p>\n
Today's the big day! Congratulations to all our high school graduates!<\/p>\n
We hope you enjoy your graduation ceremonies and wish you the best of luck in your future endeavors.<\/span><\/p>\n
Section 2. Moving to\n<\/p><\/div>\n<\/div>\n
\n
\nSection 1. High school graduation<\/p>\n
Today's the big day! Congratulations to all our high school graduates!<\/p>\n
We hope you enjoy your graduation ceremonies and wish you the best of luck in your future endeavors.<\/span><\/p>\n
Section 2. Moving to San Francisco\n<\/p><\/div>\n<\/div>\n
\n
\nSection 1. High school graduation<\/p>\n
Today's the big day! I'm graduating from high school. I'm so excited. I'm going to move to San Francisco and start a new life. I can't wait to get there.<\/span><\/p>\n
Section 2. Moving to San Francisco\n<\/p><\/div>\n<\/div>\n
\n
\nSection 1. High school graduation<\/p>\n
Today's the big day! I'm graduating from high school. I'm so excited. I'm going to move to San Francisco and start a new life. I can't wait to get there.<\/span><\/p>\n
Section 2. Moving to\n<\/p><\/div>\n<\/div>\n
\n
\nSection 1. High school graduation<\/p>\n
Today's the big day! I'm graduating from high school. I'm so excited. I'm going to move to San Francisco and start a new life. I can't wait to get there.<\/span><\/p>\n
Section 2. Moving to Istanbul\n<\/p><\/div>\n<\/div>\n
\n
\nSection 1. High school graduation<\/p>\n
Today's the big day! I'm graduating from high school. I'm so excited to be moving on to the next chapter of my life. I'm going to Istanbul!<\/span><\/p>\n
Section 2. Moving to Istanbul\n<\/p><\/div>\n<\/div>\n
\n
\nSection 1. High school graduation<\/p>\n
Today's the big day! I'm graduating from high school. I'm so excited to be moving on to the next chapter of my life. I'm going to Istanbul!<\/span><\/p>\n
Section 2. Moving to\n<\/p><\/div>\n<\/div>\n
\n
\nSection 1. High school graduation<\/p>\n
Today's the big day! I'm graduating from high school. I'm so excited to be moving on to the next chapter of my life. I'm going to Istanbul!<\/span><\/p>\n
Section 2. Moving to the\n<\/p><\/div>\n<\/div>\n
\n
\nSection 1. High school graduation<\/p>\n
Today's the big day! I'm graduating from high school. I'm so excited to be moving on to the next chapter of my life. I'm going to Istanbul!<\/span><\/p>\n
Section 2. Moving to the farm\n<\/p><\/div>\n<\/div>\n
\n
\nSection 1. High school graduation<\/p>\n
Today's the big day! My high school graduation! I'm so excited! I can't wait to leave the city and move to the farm. I want to live a simple, quiet life. I want to get up in the morning, milk the cows, feed the horses, then sit and watch the sunrise. I want to spend the rest of my days traveling the world, reading and writing. I'm so tired of this city life.<\/span><\/p>\n
Section 2. Moving to the farm\n<\/p><\/div>\n<\/div>\n<\/div>\n
\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n
In the example above, the desire is to fill-in text between two section headers of an outline. Without the context of future sections, the model generates a completion that isn't relevant to the second section. When the context of future sections are accounted for, the model generates a completion that ties the two sections together.<\/p>\n
\n
\n
\n
\n
\n
\n
def get_files(path: str, size: int):\n    def___<\/span>\n    for (dirp, _, files) in os.walk(path):\n        yield from prune(dirp, files)<\/code><\/pre>\n<\/div>\n
\n
def get_files(path: str, size: int):\n    def prune(dirp, files):\n        for file in files:\n            file = os.path.join(dirp, file)\n            if os.path.getsize(file) > size:\n                yield file<\/span>\n    for (dirp, _, files) in os.walk(path):\n        yield from prune(dirp, files)<\/code><\/pre>\n<\/div>\n
\n

def get_files(path: str, size: int):\n    def prune(dirp, files):\n        for file in files:\n            file = os.path.join(dirp, file)\n            if os.path.getsize(file) > size:\n                yield file<\/span>\n    for (dirp, _, files) in os.walk(path):\n        yield from prune(dirp, files)<\/code><\/pre>\n<\/div>\n
\n
___<\/span>\ndef get_files(path: str, size: int):\n    def prune(dirp, files):\n        for file in files:\n            file = os.path.join(dirp, file)\n            if os.path.getsize(file) > size:\n                yield file\n    for (dirp, _, files) in os.walk(path):\n        yield from prune(dirp, files)<\/code><\/pre>\n<\/div>\n
\n
import os<\/span>\ndef get_files(path: str, size: int):\n    def prune(dirp, files):\n        for file in files:\n            file = os.path.join(dirp, file)\n            if os.path.getsize(file) > size:\n                yield file\n    for (dirp, _, files) in os.walk(path):\n        yield from prune(dirp, files)<\/code><\/pre>\n<\/div>\n
\n
import os\ndef get_files(path: str, size: int):\n    """\n    def prune(dirp, files):\n        for file in files:\n            file = os.path.join(dirp, file)\n            if os.path.getsize(file) > size:\n                yield file\n    for (dirp, _, files) in os.walk(path):\n        yield from prune(dirp, files)<\/code><\/pre>\n<\/div>\n
\n
import os\ndef get_files(path: str, size: int):\n    """___<\/span>\n    def prune(dirp, files):\n        for file in files:\n            file = os.path.join(dirp, file)\n            if os.path.getsize(file) > size:\n                yield file\n    for (dirp, _, files) in os.walk(path):\n        yield from prune(dirp, files)<\/code><\/pre>\n<\/div>\n
\n
import os\ndef get_files(path: str, size: int):\n    """___<\/span>"""\n    def prune(dirp, files):\n        for file in files:\n            file = os.path.join(dirp, file)\n            if os.path.getsize(file) > size:\n                yield file\n    for (dirp, _, files) in os.walk(path):\n        yield from prune(dirp, files)<\/code><\/pre>\n<\/div>\n
\n
import os\ndef get_files(path: str, size: int):\n    """Yields files in the path tree of min size<\/span>"""\n    def prune(dirp, files):\n        for file in files:\n            file = os.path.join(dirp, file)\n            if os.path.getsize(file) > size:\n                yield file\n    for (dirp, _, files) in os.walk(path):\n        yield from prune(dirp, files)<\/code><\/pre>\n<\/div>\n<\/div>\n
\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n
Insert is particularly useful for writing code. In fact, Codex was our original motivation for developing this capability, since in software development we typically add code to the middle of an existing file where code is present before and after the completion. In the example above, the model successfully completes the missing function prune<\/code>, while connecting to code already written. We also add a docstring and missing imports, which is not possible without knowing the code that comes after. In GitHub Copilot, Insert is currently being piloted with early promising results.<\/p>\n
The insert capability is available in the API today in beta, as part of the completions endpoint<\/a> and via a new interface in Playground<\/a>. The capability can be used with the latest versions of GPT-3 and Codex, text-davinci-002<\/code> and code-davinci-002<\/code>. Pricing<\/a> is the same as previous versions of Davinci.<\/p>\n
Editing existing text<\/h2>\n
A meaningful part of writing text and code is spent editing existing content. We’ve released a new endpoint in beta called edits<\/a> that changes existing text via an instruction, instead of completing it.<\/p>\n
\n
\n
\n
\n
\n
\n
 <\/div>\n
\n<\/div>\n<\/div>\n
\n
Add<\/div>\n
\n<\/div>\n<\/div>\n
\n
Add a<\/div>\n
\n<\/div>\n<\/div>\n
\n
Add a short poem<\/div>\n
\n<\/div>\n<\/div>\n
\n
Add a short poem about<\/div>\n
\n<\/div>\n<\/div>\n
\n
Add a short poem about GPT-3<\/div>\n
\n<\/div>\n<\/div>\n
\n
Add a short poem about GPT-3<\/div>\n
\nGPT-3 is a very nice AI
\n
That's pretty good at writing replies
\n
When it's asked a question
\n
It gives its suggestion
\n
This is a poem it made that rhymes\n<\/div>\n<\/div>\n
\n
 <\/div>\n
\nGPT-3 is a very nice AI
\n
That's pretty good at writing replies
\n
When it's asked a question
\n
It gives its suggestion
\n
This is a poem it made that rhymes<\/span>\n<\/div>\n<\/div>\n
\n
Make<\/div>\n
\nGPT-3 is a very nice AI
\n
That's pretty good at writing replies
\n
When it's asked a question
\n
It gives its suggestion
\n
This is a poem it made that rhymes<\/span>\n<\/div>\n<\/div>\n
\n
Make this<\/div>\n
\nGPT-3 is a very nice AI
\n
That's pretty good at writing replies
\n
When it's asked a question
\n
It gives its suggestion
\n
This is a poem it made that rhymes<\/span>\n<\/div>\n<\/div>\n
\n
Make this in the<\/div>\n
\nGPT-3 is a very nice AI
\n
That's pretty good at writing replies
\n
When it's asked a question
\n
It gives its suggestion
\n
This is a poem it made that rhymes<\/span>\n<\/div>\n<\/div>\n
\n
Make this in the voice<\/div>\n
\nGPT-3 is a very nice AI
\n
That's pretty good at writing replies
\n
When it's asked a question
\n
It gives its suggestion
\n
This is a poem it made that rhymes<\/span>\n<\/div>\n<\/div>\n
\n
Make this in the voice of GPT-3<\/div>\n
\nGPT-3 is a very nice AI
\n
That's pretty good at writing replies
\n
When it's asked a question
\n
It gives its suggestion
\n
This is a poem it made that rhymes<\/span>\n<\/div>\n<\/div>\n
\n
Make this in the voice of GPT-3<\/div>\n
\nI am a very nice AI
\n
I am pretty good at writing replies
\n
When I am asked a question
\n
I give my suggestion
\n
This is a poem it made that rhymes\n<\/div>\n<\/div>\n
\n
 <\/div>\n
\nI am a very nice AI
\n
I am pretty good at writing replies
\n
When I am asked a question
\n
I give my suggestion
\n
This is a poem it made that rhymes<\/span>\n<\/div>\n<\/div>\n
\n
Format<\/div>\n
\nI am a very nice AI
\n
I am pretty good at writing replies
\n
When I am asked a question
\n
I give my suggestion
\n
This is a poem it made that rhymes<\/span>\n<\/div>\n<\/div>\n
\n
Format this<\/div>\n
\nI am a very nice AI
\n
I am pretty good at writing replies
\n
When I am asked a question
\n
I give my suggestion
\n
This is a poem it made that rhymes<\/span>\n<\/div>\n<\/div>\n
\n
Format this like a<\/div>\n
\nI am a very nice AI
\n
I am pretty good at writing replies
\n
When I am asked a question
\n
I give my suggestion
\n
This is a poem it made that rhymes<\/span>\n<\/div>\n<\/div>\n
\n
Format this like a letter<\/div>\n
\nI am a very nice AI
\n
I am pretty good at writing replies
\n
When I am asked a question
\n
I give my suggestion
\n
This is a poem it made that rhymes<\/span>\n<\/div>\n<\/div>\n
\n
Format this like a letter and sign<\/div>\n
\nI am a very nice AI
\n
I am pretty good at writing replies
\n
When I am asked a question
\n
I give my suggestion
\n
This is a poem it made that rhymes<\/span>\n<\/div>\n<\/div>\n
\n
Format this like a letter and sign from GPT-3<\/div>\n
\nI am a very nice AI
\n
I am pretty good at writing replies
\n
When I am asked a question
\n
I give my suggestion
\n
This is a poem it made that rhymes<\/span>\n<\/div>\n<\/div>\n
\n
Format this like a letter and sign from GPT-3<\/div>\n
\nDear Human,<\/p>\n
I am very nice AI
\n
I am pretty good at writing replies
\n
When I am asked a question
\n
I give my suggestion
\n
This is a poem I made that rhymes<\/p>\n
Kind regards,
\n
GPT-3\n<\/div>\n<\/div>\n<\/div>\n
\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n
Editing works by specifying existing text as a prompt and an instruction on how to modify it. The edits endpoint can be used to change the tone or structure of text, or make targeted changes like fixing spelling. We’ve also observed edits to work well on empty prompts, thus enabling text generation similar to the completions endpoint<\/a>. In the example above, we use edits to (1) add a poem, (2) change the poem to be in first-person, (3) transform the poem into a letter, with the appropriate salutation and signature.<\/p>\n
\n
\n
\n
\n
\n
\n
 <\/div>\n
The three US cities with the worst traffic are:\n
1. Boston, MA (164 hours)\n2. Washington, DC (155 hours)\n3. Chicago, IL (138 hours)<\/span><\/code><\/pre>\n<\/div>\n
\n
Convert<\/div>\n
The three US cities with the worst traffic are:\n
1. Boston, MA (164 hours)\n2. Washington, DC (155 hours)\n3. Chicago, IL (138 hours)<\/span><\/code><\/pre>\n<\/div>\n
\n
Convert the list<\/div>\n
The three US cities with the worst traffic are:\n
1. Boston, MA (164 hours)\n2. Washington, DC (155 hours)\n3. Chicago, IL (138 hours)<\/span><\/code><\/pre>\n<\/div>\n
\n
Convert the list to JSONL file<\/div>\n
The three US cities with the worst traffic are:\n
1. Boston, MA (164 hours)\n2. Washington, DC (155 hours)\n3. Chicago, IL (138 hours)<\/span><\/code><\/pre>\n<\/div>\n
\n
Convert the list to JSONL file with rank<\/div>\n
The three US cities with the worst traffic are:\n
1. Boston, MA (164 hours)\n2. Washington, DC (155 hours)\n3. Chicago, IL (138 hours)<\/span><\/code><\/pre>\n<\/div>\n
\n
Convert the list to JSONL file with rank, city<\/div>\n
The three US cities with the worst traffic are:\n
1. Boston, MA (164 hours)\n2. Washington, DC (155 hours)\n3. Chicago, IL (138 hours)<\/span><\/code><\/pre>\n<\/div>\n
\n
Convert the list to JSONL file with rank, city, state<\/div>\n
The three US cities with the worst traffic are:\n
1. Boston, MA (164 hours)\n2. Washington, DC (155 hours)\n3. Chicago, IL (138 hours)<\/span><\/code><\/pre>\n<\/div>\n
\n
Convert the list to JSONL file with rank, city, state, and hours<\/div>\n
The three US cities with the worst traffic are:\n
1. Boston, MA (164 hours)\n2. Washington, DC (155 hours)\n3. Chicago, IL (138 hours)<\/span><\/code><\/pre>\n<\/div>\n
\n
Convert the list to JSONL file with rank, city, state, and hours<\/div>\n
[\n  {"rank": 1, "city": "Boston", "state": "MA", "hours": 164},\n  {"rank": 2, "city": "Washington DC", "state": "DC", "hours": 155},\n  {"rank": 3, "city": "Chicago", "state": "IL", "hours": 138}\n]<\/code><\/pre>\n<\/div>\n
\n
 <\/div>\n
[\n  {"rank": 1, "city": "Boston", "state": "MA", "hours": 164},\n  {"rank": 2, "city": "Washington DC", "state": "DC", "hours": 155},\n  {"rank": 3, "city": "Chicago", "state": "IL", "hours": 138}\n]<\/span><\/code><\/pre>\n<\/div>\n
\n
Remove<\/div>\n
[\n  {"rank": 1, "city": "Boston", "state": "MA", "hours": 164},\n  {"rank": 2, "city": "Washington DC", "state": "DC", "hours": 155},\n  {"rank": 3, "city": "Chicago", "state": "IL", "hours": 138}\n]<\/span><\/code><\/pre>\n<\/div>\n
\n
Remove rank<\/div>\n
[\n  {"rank": 1, "city": "Boston", "state": "MA", "hours": 164},\n  {"rank": 2, "city": "Washington DC", "state": "DC", "hours": 155},\n  {"rank": 3, "city": "Chicago", "state": "IL", "hours": 138}\n]<\/span><\/code><\/pre>\n<\/div>\n
\n
Remove rank and write out<\/div>\n
[\n  {"rank": 1, "city": "Boston", "state": "MA", "hours": 164},\n  {"rank": 2, "city": "Washington DC", "state": "DC", "hours": 155},\n  {"rank": 3, "city": "Chicago", "state": "IL", "hours": 138}\n]<\/span><\/code><\/pre>\n<\/div>\n
\n
Remove rank and write out full state names<\/div>\n
[\n  {"rank": 1, "city": "Boston", "state": "MA", "hours": 164},\n  {"rank": 2, "city": "Washington DC", "state": "DC", "hours": 155},\n  {"rank": 3, "city": "Chicago", "state": "IL", "hours": 138}\n]<\/span><\/code><\/pre>\n<\/div>\n
\n
Remove rank and write out full state names<\/div>\n
[\n  {"city": "Boston", "state": "Massachusetts", "hours": 164},\n  {"city": "Washington DC", "state": "District of Columbia", "hours": 155},\n  {"city": "Chicago", "state": "Illinois", "hours": 138}\n]<\/code><\/pre>\n<\/div>\n
\n
 <\/div>\n
[\n  {"city": "Boston", "state": "Massachusetts", "hours": 164},\n  {"city": "Washington DC", "state": "District of Columbia", "hours": 155},\n  {"city": "Chicago", "state": "Illinois", "hours": 138}\n]<\/span><\/code><\/pre>\n<\/div>\n
\n
Turn<\/div>\n
[\n  {"city": "Boston", "state": "Massachusetts", "hours": 164},\n  {"city": "Washington DC", "state": "District of Columbia", "hours": 155},\n  {"city": "Chicago", "state": "Illinois", "hours": 138}\n]<\/span><\/code><\/pre>\n<\/div>\n
\n
Turn into YAML<\/div>\n
[\n  {"city": "Boston", "state": "Massachusetts", "hours": 164},\n  {"city": "Washington DC", "state": "District of Columbia", "hours": 155},\n  {"city": "Chicago", "state": "Illinois", "hours": 138}\n]<\/span><\/code><\/pre>\n<\/div>\n
\n
Turn into YAML and return<\/div>\n
[\n  {"city": "Boston", "state": "Massachusetts", "hours": 164},\n  {"city": "Washington DC", "state": "District of Columbia", "hours": 155},\n  {"city": "Chicago", "state": "Illinois", "hours": 138}\n]<\/span><\/code><\/pre>\n<\/div>\n
\n
Turn into YAML and return from a function<\/div>\n
[\n  {"city": "Boston", "state": "Massachusetts", "hours": 164},\n  {"city": "Washington DC", "state": "District of Columbia", "hours": 155},\n  {"city": "Chicago", "state": "Illinois", "hours": 138}\n]<\/span><\/code><\/pre>\n<\/div>\n
\n
Turn into YAML and return from a function called get_yaml<\/div>\n
[\n  {"city": "Boston", "state": "Massachusetts", "hours": 164},\n  {"city": "Washington DC", "state": "District of Columbia", "hours": 155},\n  {"city": "Chicago", "state": "Illinois", "hours": 138}\n]<\/span><\/code><\/pre>\n<\/div>\n
\n
Turn into YAML and return from a function called get_yaml<\/div>\n
def get_yaml():\n    return """\n    - city: Boston\n      state: Massachusetts\n      hours: 164\n    - city: Washington DC\n      state: District of Columbia\n      hours: 155\n    - city: Chicago\n      state: Illinois\n      hours: 138\n    """<\/code><\/pre>\n<\/div>\n<\/div>\n
\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n
The edits endpoint is particularly useful for writing code. It works well for tasks like refactoring, adding documentation, translating between programming languages, and changing coding style. The example above starts with JSON input containing cities ranked by population. With our first edit, Codex removes the rank field from the JSON, and changes the state abbreviations into full names. The second edit converts the JSON file into YAML returned from a function.<\/p>\n
Editing is available as a specialized endpoint in the API<\/a> and through a new interface in Playground<\/a>. It is supported by models text-davinci-edits-001<\/code> and code-davinci-edits-001<\/code>. The edits endpoint is currently free to use<\/a> and publicly available as a beta.<\/p>\n