{
"$type": "site.standard.document",
"bskyPostRef": {
"cid": "bafyreiaxtk32m4md4lscyf3osk3bqw5tidorxowtn7uhlx54ssdets7t4q",
"uri": "at://did:plc:pgryn3ephfd2xgft23qokfzt/app.bsky.feed.post/3mpswqulb5xw2"
},
"path": "/t/how-to-publish-my-research-in-huggingface/177341#post_6",
"publishedAt": "2026-07-04T09:44:47.000Z",
"site": "https://discuss.huggingface.co",
"tags": [
"Presenting TIS (Token Importance Scoring) - A new way to compress KV cache",
"Research"
],
"textContent": "And it’s alive!\n\nPresenting TIS (Token Importance Scoring) - A new way to compress KV cache Research\n\n> ## Token Importance Scoring for KV Cache Compression I have spent some time experimenting with learned token importance for efficient KV cache compression in LLMs. The result is a simple mechanism that works surprisingly well, especially on synthetic retrieval tasks. ### What’s This About Most KV cache compression methods rely on static heuristics (like position-based selection). I asked myself: what if we just learned which tokens matter. It turns out constraint-aware learning is the key - h…",
"title": "How to publish my research in HuggingFace?"
}