transformers/main/en/bertology.html

<meta charset="utf-8" /><meta name="hf:doc:metadata" content="{&quot;title&quot;:&quot;BERTology&quot;,&quot;local&quot;:&quot;bertology&quot;,&quot;sections&quot;:[],&quot;depth&quot;:1}">
		<link href="/docs/transformers/main/en/_app/immutable/assets/0.e3b0c442.css" rel="modulepreload">
		<link rel="modulepreload" href="/docs/transformers/main/en/_app/immutable/entry/start.2135b7e6.js">
		<link rel="modulepreload" href="/docs/transformers/main/en/_app/immutable/chunks/scheduler.25b97de1.js">
		<link rel="modulepreload" href="/docs/transformers/main/en/_app/immutable/chunks/singletons.0f2b7d5f.js">
		<link rel="modulepreload" href="/docs/transformers/main/en/_app/immutable/chunks/index.e188933d.js">
		<link rel="modulepreload" href="/docs/transformers/main/en/_app/immutable/chunks/paths.3d04d2c6.js">
		<link rel="modulepreload" href="/docs/transformers/main/en/_app/immutable/entry/app.24372c84.js">
		<link rel="modulepreload" href="/docs/transformers/main/en/_app/immutable/chunks/index.d9030fc9.js">
		<link rel="modulepreload" href="/docs/transformers/main/en/_app/immutable/nodes/0.026d2fdd.js">
		<link rel="modulepreload" href="/docs/transformers/main/en/_app/immutable/chunks/each.e59479a4.js">
		<link rel="modulepreload" href="/docs/transformers/main/en/_app/immutable/nodes/10.3b530d34.js">
		<link rel="modulepreload" href="/docs/transformers/main/en/_app/immutable/chunks/EditOnGithub.91d95064.js"><!-- HEAD_svelte-u9bgzb_START --><meta name="hf:doc:metadata" content="{&quot;title&quot;:&quot;BERTology&quot;,&quot;local&quot;:&quot;bertology&quot;,&quot;sections&quot;:[],&quot;depth&quot;:1}"><!-- HEAD_svelte-u9bgzb_END -->      <p></p>   <h1 class="relative group"><a id="bertology" class="header-link block pr-1.5 text-lg no-hover:hidden with-hover:absolute with-hover:p-1.5 with-hover:opacity-0 with-hover:group-hover:opacity-100 with-hover:right-full" href="#bertology"><span><svg class="" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" aria-hidden="true" role="img" width="1em" height="1em" preserveAspectRatio="xMidYMid meet" viewBox="0 0 256 256"><path d="M167.594 88.393a8.001 8.001 0 0 1 0 11.314l-67.882 67.882a8 8 0 1 1-11.314-11.315l67.882-67.881a8.003 8.003 0 0 1 11.314 0zm-28.287 84.86l-28.284 28.284a40 40 0 0 1-56.567-56.567l28.284-28.284a8 8 0 0 0-11.315-11.315l-28.284 28.284a56 56 0 0 0 79.196 79.197l28.285-28.285a8 8 0 1 0-11.315-11.314zM212.852 43.14a56.002 56.002 0 0 0-79.196 0l-28.284 28.284a8 8 0 1 0 11.314 11.314l28.284-28.284a40 40 0 0 1 56.568 56.567l-28.285 28.285a8 8 0 0 0 11.315 11.314l28.284-28.284a56.065 56.065 0 0 0 0-79.196z" fill="currentColor"></path></svg></span></a> <span>BERTology</span></h1> <p data-svelte-h="svelte-e5vxoh">There is a growing field of study concerned with investigating the inner working of large-scale transformers like BERT
(that some call “BERTology”). Some good examples of this field are:</p> <ul data-svelte-h="svelte-1e34lw"><li>BERT Rediscovers the Classical NLP Pipeline by Ian Tenney, Dipanjan Das, Ellie Pavlick:
<a href="https://arxiv.org/abs/1905.05950" rel="nofollow">https://arxiv.org/abs/1905.05950</a></li> <li>Are Sixteen Heads Really Better than One? by Paul Michel, Omer Levy, Graham Neubig: <a href="https://arxiv.org/abs/1905.10650" rel="nofollow">https://arxiv.org/abs/1905.10650</a></li> <li>What Does BERT Look At? An Analysis of BERT’s Attention by Kevin Clark, Urvashi Khandelwal, Omer Levy, Christopher D.
Manning: <a href="https://arxiv.org/abs/1906.04341" rel="nofollow">https://arxiv.org/abs/1906.04341</a></li> <li>CAT-probing: A Metric-based Approach to Interpret How Pre-trained Models for Programming Language Attend Code Structure: <a href="https://arxiv.org/abs/2210.04633" rel="nofollow">https://arxiv.org/abs/2210.04633</a></li></ul> <p data-svelte-h="svelte-1bavv9a">In order to help this new field develop, we have included a few additional features in the BERT/GPT/GPT-2 models to
help people access the inner representations, mainly adapted from the great work of Paul Michel
(<a href="https://arxiv.org/abs/1905.10650" rel="nofollow">https://arxiv.org/abs/1905.10650</a>):</p> <ul data-svelte-h="svelte-14erpmf"><li>accessing all the hidden-states of BERT/GPT/GPT-2,</li> <li>accessing all the attention weights for each head of BERT/GPT/GPT-2,</li> <li>retrieving heads output values and gradients to be able to compute head importance score and prune head as explained
in <a href="https://arxiv.org/abs/1905.10650" rel="nofollow">https://arxiv.org/abs/1905.10650</a>.</li></ul> <p data-svelte-h="svelte-2lqgzt">To help you understand and use these features, we have added a specific example script: <a href="https://github.com/huggingface/transformers/tree/main/examples/research_projects/bertology/run_bertology.py" rel="nofollow">bertology.py</a> which extracts information and prune a model pre-trained on
GLUE.</p> <a class="!text-gray-400 !no-underline text-sm flex items-center not-prose mt-4" href="https://github.com/huggingface/transformers/blob/main/docs/source/en/bertology.md" target="_blank"><span data-svelte-h="svelte-1kd6by1">&lt;</span> <span data-svelte-h="svelte-x0xyl0">&gt;</span> <span data-svelte-h="svelte-1dajgef"><span class="underline ml-1.5">Update</span> on GitHub</span></a>  <p></p> 
			
			<script>
				{
					__sveltekit_1xexzbk = {
						assets: "/docs/transformers/main/en",
						base: "/docs/transformers/main/en",
						env: {}
					};

					const element = document.currentScript.parentElement;

					const data = [null,null];

					Promise.all([
						import("/docs/transformers/main/en/_app/immutable/entry/start.2135b7e6.js"),
						import("/docs/transformers/main/en/_app/immutable/entry/app.24372c84.js")
					]).then(([kit, app]) => {
						kit.start(app, element, {
							node_ids: [0, 10],
							data,
							form: null,
							error: null
						});
					});
				}
			</script>