Viewing enriched Twitter post
I've tested the "long is more" trick on @Teknium1's OpenHermes dataset and it works surprisingly well 🔥! - Select the 1k longest samples (0.1%) - SFT Mistral-7B for 15 epochs with NEFTune α=10 - MT Bench ~7 + decent perf on other benchmarks 💾Dataset: https://t.co/MWnWXgZ6QT https://t.co/0HpXvX1DAm
{
"media": [
{
"id": "",
"type": "photo",
"url": null,
"media_url": "https://pbs.twimg.com/media/GGd9AWcWsAEpf7U.png",
"media_url_https": null,
"display_url": null,
"expanded_url": null
}
],
"nlp": {
"sentiment": "positive",
"processed_at": "2025-08-06T12:51:06.346153"
},
"score": 0.94,
"scored_at": "2025-08-09T13:46:07.544797",
"import_source": "manual_curation_2024",
"score_components": {
"author": 0.09,
"engagement": 0.11891989504740688,
"quality": 0.16000000000000003,
"source": 0.15,
"nlp": 0.1,
"recency": 0.010000000000000002
},
"source_tagged_at": "2025-08-09T13:43:02.898905",
"enriched": true,
"enriched_at": "2025-08-09T13:43:02.898907",
"links_checked": true,
"checked_at": "2025-08-10T10:32:33.766977",
"original_structure": "had_media_only"
}{
"user": {
"created_at": "2018-08-14T22:21:16.000Z",
"default_profile_image": false,
"description": "🤗 Putting hugs in RLHF @huggingface\n📖 Co-author of \"NLP with Transformers\" book\n💥 Ex-particle physicist\n🤘 Occasional guitarist\n🇦🇺 in 🇨🇭",
"fast_followers_count": 0,
"favourites_count": 8128,
"followers_count": 8175,
"friends_count": 401,
"has_custom_timelines": true,
"is_translator": false,
"listed_count": 211,
"location": "Berne, Switzerland",
"media_count": 597,
"name": "Lewis Tunstall",
"normal_followers_count": 8175,
"possibly_sensitive": false,
"profile_banner_url": "https://pbs.twimg.com/profile_banners/1029493180704714753/1655469477",
"profile_image_url_https": "https://pbs.twimg.com/profile_images/1097405296543236096/gS2C7RIq_normal.jpg",
"screen_name": "_lewtun",
"statuses_count": 2887,
"translator_type": "none",
"url": "https://t.co/F3W4xU7x2Z",
"verified": false,
"withheld_in_countries": [],
"id_str": "1029493180704714753"
},
"id": "1758520258132865210",
"conversation_id": "1758520258132865210",
"full_text": "I've tested the \"long is more\" trick on @Teknium1's OpenHermes dataset and it works surprisingly well 🔥!\n\n- Select the 1k longest samples (0.1%)\n- SFT Mistral-7B for 15 epochs with NEFTune α=10\n- MT Bench ~7 + decent perf on other benchmarks\n\n💾Dataset: https://t.co/MWnWXgZ6QT https://t.co/0HpXvX1DAm",
"reply_count": 6,
"retweet_count": 24,
"favorite_count": 190,
"hashtags": [],
"symbols": [],
"user_mentions": [
{
"id_str": "1365020011123773442",
"name": "Teknium (e/λ)",
"screen_name": "Teknium1",
"profile": "https://twitter.com/Teknium1"
}
],
"urls": [
{
"url": "https://t.co/MWnWXgZ6QT",
"expanded_url": "https://huggingface.co/datasets/HuggingFaceH4/OpenHermes-2.5-1k-longest",
"display_url": "huggingface.co/datasets/Huggi…"
}
],
"media": [
{
"media_url": "https://pbs.twimg.com/media/GGd9AWcWsAEpf7U.png",
"type": "photo"
}
],
"url": "https://twitter.com/_lewtun/status/1758520258132865210",
"created_at": "2024-02-16T15:54:35.000Z",
"#sort_index": "1758520258132865210",
"view_count": 20889,
"quote_count": 1,
"is_quote_tweet": false,
"is_retweet": false,
"is_pinned": false,
"is_truncated": false,
"startUrl": "https://twitter.com/_lewtun/status/1758520258132865210"
}