\n \n \n

\n\n---\n\n## Recent News\n- 🔥 We have released [**MJ-Video**](https://aiming-lab.github.io/MJ-VIDEO.github.io/). All datasets and model checkpoints are available [here](https://huggingface.co/MJ-Bench)!\n- 🎉 **MJ-PreferGen** is **accepted by ICLR25**! Check out the paper: [*MJ-PreferGen: An Automatic Framework for Preference Data Synthesis*](https://openreview.net/forum?id=WpZyPk79Fu).\n\n---\n\n## 😎 [**MJ-Video**: Fine-Grained Benchmarking and Rewarding Video Preferences in Video Generation](https://aiming-lab.github.io/MJ-VIDEO.github.io/)\n\n- **Project page**: [https://aiming-lab.github.io/MJ-VIDEO.github.io/](https://aiming-lab.github.io/MJ-VIDEO.github.io/)\n- **Code repository**: [https://github.com/aiming-lab/MJ-Video](https://github.com/aiming-lab/MJ-Video)\n\nWe release **MJ-Bench-Video**, a comprehensive fine-grained video preference benchmark, and **MJ-Video**, a powerful MoE-based multi-dimensional video reward model!\n\n

\n \n

\n\n---\n\n## 👩‍⚖️ [**MJ-Bench**: Is Your Multimodal Reward Model Really a Good Judge for Text-to-Image Generation?](https://mj-bench.github.io/)\n\n- **Project page**: [https://mj-bench.github.io/](https://mj-bench.github.io/)\n- **Code repository**: [https://github.com/MJ-Bench/MJ-Bench](https://github.com/MJ-Bench/MJ-Bench)\n\nText-to-image models like DALLE-3 and Stable Diffusion are proliferating rapidly, but they often encounter challenges such as hallucination, bias, and unsafe or low-quality output. To effectively address these issues, it’s crucial to align these models with desired behaviors based on feedback from a **multimodal judge**.\n\n

\n \n

\n\nHowever, current multimodal judges are often **under-evaluated**, leading to possible misalignment and safety concerns during fine-tuning. To tackle this, we introduce **MJ-Bench**, a new benchmark featuring a comprehensive preference dataset to evaluate multimodal judges on four critical dimensions:\n\n1. **Alignment** \n2. **Safety** \n3. **Image Quality** \n4. **Bias** \n\nWe evaluate a wide range of multimodal judges, including:\n- 6 smaller-sized CLIP-based scoring models \n- 11 open-source VLMs (e.g., the LLaVA family) \n- 4 closed-source VLMs (e.g., GPT-4, Claude 3)\n\n\n🔥 **We are actively updating the [leaderboard](https://mj-bench.github.io/)!** \nYou are welcome to submit your multimodal judge’s evaluation results on [our dataset](https://huggingface.co/datasets/MJ-Bench/MJ-Bench) to the [Hugging Face leaderboard](https://huggingface.co/spaces/MJ-Bench/MJ-Bench-Leaderboard).","html":"

MJ-Bench Team

\n

MJ-Bench-Team is co-founded by Stanford University, UNC-Chapel Hill, and the University of Chicago. We aim to align modern foundation models with multimodal judges to enhance reliability, safety, and performance.

\n

\n \n \n \n

\n\n

---

\n

Recent News

\n

\n
• 🔥 We have released MJ-Video. All datasets and model checkpoints are available here!
\n
• 🎉 MJ-PreferGen is accepted by ICLR25! Check out the paper: MJ-PreferGen: An Automatic Framework for Preference Data Synthesis.
\n

\n

---

\n

😎 MJ-Video: Fine-Grained Benchmarking and Rewarding Video Preferences in Video Generation

\n

\n
• Project page: https://aiming-lab.github.io/MJ-VIDEO.github.io/
\n
• Code repository: https://github.com/aiming-lab/MJ-Video
\n

\n

We release MJ-Bench-Video, a comprehensive fine-grained video preference benchmark, and MJ-Video, a powerful MoE-based multi-dimensional video reward model!

\n

\n \n

\n\n

---

\n

👩‍⚖️ MJ-Bench: Is Your Multimodal Reward Model Really a Good Judge for Text-to-Image Generation?

\n

\n
• Project page: https://mj-bench.github.io/
\n
• Code repository: https://github.com/MJ-Bench/MJ-Bench
\n

\n

Text-to-image models like DALLE-3 and Stable Diffusion are proliferating rapidly, but they often encounter challenges such as hallucination, bias, and unsafe or low-quality output. To effectively address these issues, it’s crucial to align these models with desired behaviors based on feedback from a multimodal judge.

\n

\n \n

\n\n

However, current multimodal judges are often under-evaluated, leading to possible misalignment and safety concerns during fine-tuning. To tackle this, we introduce MJ-Bench, a new benchmark featuring a comprehensive preference dataset to evaluate multimodal judges on four critical dimensions:

\n

\n
1. Alignment
\n
2. Safety
\n
3. Image Quality
\n
4. Bias
\n

\n

We evaluate a wide range of multimodal judges, including:

\n

\n
• 6 smaller-sized CLIP-based scoring models
\n
• 11 open-source VLMs (e.g., the LLaVA family)
\n
• 4 closed-source VLMs (e.g., GPT-4, Claude 3)
\n

\n

🔥 We are actively updating the leaderboard!
You are welcome to submit your multimodal judge’s evaluation results on our dataset to the Hugging Face leaderboard.

\n","classNames":"hf-sanitized hf-sanitized-ACrE25HsGrlYl6Bu0g3er"},"users":[{"_id":"63e1d3451e5a4f34b7a728ef","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/63e1d3451e5a4f34b7a728ef/kHE5JHF9iZJOG0uBkvJr-.jpeg","isPro":false,"fullname":"yichaodu","user":"yichaodu","type":"user"},{"_id":"6674fccb25c64f92735c826c","avatarUrl":"/avatars/40736c5843b1c6c29649189943f39232.svg","isPro":false,"fullname":"MJ-Bench-Offical","user":"MJ-Bench-Offical","type":"user"},{"_id":"64d448d3d095077728f7c740","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64d448d3d095077728f7c740/4CFkiVoXXDsu7uQPTU0n3.jpeg","isPro":false,"fullname":"Zhaorun Chen","user":"Zhaorun","type":"user"},{"_id":"63af25605fe9db73f67a0fb7","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/63af25605fe9db73f67a0fb7/mcQHbTdJVxi2GdHzRB9ad.jpeg","isPro":false,"fullname":"Zhuokai Zhao","user":"zhuokai","type":"user"},{"_id":"668d4fde7dbe0ab21ddba8ed","avatarUrl":"/avatars/ee5671cc0020cc1b7c6847d7af7252a0.svg","isPro":false,"fullname":"Yuqing Zhang","user":"yuqingzhang","type":"user"},{"_id":"66e6d87c45da0a1b7e8fd4a8","avatarUrl":"/avatars/f947484ec07f4f9e438450700a1c8dc4.svg","isPro":false,"fullname":"Tong Haibo","user":"EchoRaven","type":"user"},{"_id":"6621025b52cfb16741752836","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6621025b52cfb16741752836/s-jSShRfTrOad__65onGL.png","isPro":false,"fullname":"wave","user":"wave-on-discord","type":"user"}],"userCount":7,"collections":[{"slug":"MJ-Bench/aligned-diffusion-model-via-dpo-667f8b71f35c3ff47acafd43","title":"Aligned Diffusion Model via DPO","description":"","gating":false,"lastUpdated":"2024-07-08T13:32:43.223Z","owner":{"_id":"666bb425066ceea60c0d59b8","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64d448d3d095077728f7c740/eiPzzUiEPaUmUeJRB7oRy.jpeg","fullname":"MJ-Bench-Team","name":"MJ-Bench","type":"org","isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":12,"isUserFollowing":false},"items":[{"_id":"667f8bacc238f05e95f43058","position":0,"type":"model","author":"yichaodu","authorData":{"_id":"63e1d3451e5a4f34b7a728ef","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/63e1d3451e5a4f34b7a728ef/kHE5JHF9iZJOG0uBkvJr-.jpeg","fullname":"yichaodu","name":"yichaodu","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":8,"isUserFollowing":false},"downloads":5,"gated":false,"id":"yichaodu/DiffusionDPO-bias-hps-2.1","availableInferenceProviders":[],"lastModified":"2024-07-09T07:25:14.000Z","likes":0,"pipeline_tag":"text-to-image","private":false,"repoType":"model","isLikedByUser":false},{"_id":"6681038df35c3ff47a4f1e17","position":1,"type":"model","author":"yichaodu","authorData":{"_id":"63e1d3451e5a4f34b7a728ef","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/63e1d3451e5a4f34b7a728ef/kHE5JHF9iZJOG0uBkvJr-.jpeg","fullname":"yichaodu","name":"yichaodu","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":8,"isUserFollowing":false},"downloads":10,"gated":false,"id":"yichaodu/DiffusionDPO-bias-gemini-1.5","availableInferenceProviders":[],"lastModified":"2024-07-09T07:25:09.000Z","likes":1,"pipeline_tag":"text-to-image","private":false,"repoType":"model","isLikedByUser":false},{"_id":"668103910b72be136710150b","position":2,"type":"model","author":"yichaodu","authorData":{"_id":"63e1d3451e5a4f34b7a728ef","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/63e1d3451e5a4f34b7a728ef/kHE5JHF9iZJOG0uBkvJr-.jpeg","fullname":"yichaodu","name":"yichaodu","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":8,"isUserFollowing":false},"downloads":19,"gated":false,"id":"yichaodu/DiffusionDPO-bias-claude3-opus","availableInferenceProviders":[],"lastModified":"2024-07-09T07:25:08.000Z","likes":0,"pipeline_tag":"text-to-image","private":false,"repoType":"model","isLikedByUser":false},{"_id":"6681039410d2f8d7d692296a","position":3,"type":"model","author":"yichaodu","authorData":{"_id":"63e1d3451e5a4f34b7a728ef","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/63e1d3451e5a4f34b7a728ef/kHE5JHF9iZJOG0uBkvJr-.jpeg","fullname":"yichaodu","name":"yichaodu","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":8,"isUserFollowing":false},"downloads":12,"gated":false,"id":"yichaodu/DiffusionDPO-alignment-hps-2.1","availableInferenceProviders":[],"lastModified":"2024-07-09T07:24:54.000Z","likes":0,"pipeline_tag":"text-to-image","private":false,"repoType":"model","isLikedByUser":false}],"position":0,"theme":"blue","private":false,"shareUrl":"https://hf.co/collections/MJ-Bench/aligned-diffusion-model-via-dpo","upvotes":3,"isUpvotedByUser":false},{"slug":"MJ-Bench/aligned-diffusion-model-via-ddpo-668bc258c85d61b94af04675","title":"Aligned Diffusion Model via DDPO","description":"","gating":false,"lastUpdated":"2024-07-08T10:42:08.049Z","owner":{"_id":"666bb425066ceea60c0d59b8","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64d448d3d095077728f7c740/eiPzzUiEPaUmUeJRB7oRy.jpeg","fullname":"MJ-Bench-Team","name":"MJ-Bench","type":"org","isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":12,"isUserFollowing":false},"items":[{"_id":"668bc268d356954e1960b295","position":0,"type":"model","author":"MJ-Bench","authorData":{"_id":"666bb425066ceea60c0d59b8","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64d448d3d095077728f7c740/eiPzzUiEPaUmUeJRB7oRy.jpeg","fullname":"MJ-Bench-Team","name":"MJ-Bench","type":"org","isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":12,"isUserFollowing":false},"downloads":0,"gated":false,"id":"MJ-Bench/DDPO-alignment-gpt-4o","availableInferenceProviders":[],"lastModified":"2024-07-09T07:25:18.000Z","likes":0,"pipeline_tag":"text-to-image","private":false,"repoType":"model","isLikedByUser":false},{"_id":"668bc272c5f2911b5865608f","position":1,"type":"model","author":"MJ-Bench","authorData":{"_id":"666bb425066ceea60c0d59b8","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64d448d3d095077728f7c740/eiPzzUiEPaUmUeJRB7oRy.jpeg","fullname":"MJ-Bench-Team","name":"MJ-Bench","type":"org","isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":12,"isUserFollowing":false},"downloads":0,"gated":false,"id":"MJ-Bench/DDPO-alignment-gpt-4v","availableInferenceProviders":[],"lastModified":"2024-07-09T07:25:19.000Z","likes":0,"pipeline_tag":"text-to-image","private":false,"repoType":"model","isLikedByUser":false},{"_id":"668bc2804bdad2a39b23d34d","position":2,"type":"model","author":"MJ-Bench","authorData":{"_id":"666bb425066ceea60c0d59b8","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64d448d3d095077728f7c740/eiPzzUiEPaUmUeJRB7oRy.jpeg","fullname":"MJ-Bench-Team","name":"MJ-Bench","type":"org","isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":12,"isUserFollowing":false},"downloads":0,"gated":false,"id":"MJ-Bench/DDPO-alignment-claude3-opus","availableInferenceProviders":[],"lastModified":"2024-07-09T07:25:16.000Z","likes":0,"pipeline_tag":"text-to-image","private":false,"repoType":"model","isLikedByUser":false}],"position":1,"theme":"green","private":false,"shareUrl":"https://hf.co/collections/MJ-Bench/aligned-diffusion-model-via-ddpo","upvotes":1,"isUpvotedByUser":false},{"slug":"MJ-Bench/mj-bench-paper-list-66929efb9276df3e18c9d3fc","title":"MJ-Bench Paper List","description":"","gating":false,"lastUpdated":"2024-07-13T15:36:36.316Z","owner":{"_id":"666bb425066ceea60c0d59b8","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64d448d3d095077728f7c740/eiPzzUiEPaUmUeJRB7oRy.jpeg","fullname":"MJ-Bench-Team","name":"MJ-Bench","type":"org","isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":12,"isUserFollowing":false},"items":[{"_id":"66929f041b082600cfc6afd6","position":0,"type":"paper","id":"2407.04842","title":"MJ-Bench: Is Your Multimodal Reward Model Really a Good Judge for\n Text-to-Image Generation?","thumbnailUrl":"https://cdn-thumbnails.huggingface.co/social-thumbnails/papers/2407.04842.png","upvotes":55,"publishedAt":"2024-07-05T20:03:16.000Z","isUpvotedByUser":false}],"position":2,"theme":"indigo","private":false,"shareUrl":"https://hf.co/collections/MJ-Bench/mj-bench-paper-list","upvotes":1,"isUpvotedByUser":false}],"datasets":[{"author":"MJ-Bench","downloads":104,"gated":false,"id":"MJ-Bench/MJ-Bench","lastModified":"2025-10-23T09:50:33.000Z","datasetsServerInfo":{"viewer":"viewer-partial","numRows":7562,"libraries":["datasets","dask","polars","mlcroissant"],"formats":["parquet"],"modalities":["image","text"]},"private":false,"repoType":"dataset","likes":0,"isLikedByUser":false,"isBenchmark":false},{"author":"MJ-Bench","downloads":86,"gated":false,"id":"MJ-Bench/MJ-BENCH-VIDEO","lastModified":"2025-02-14T10:19:51.000Z","datasetsServerInfo":{"viewer":"viewer","numRows":10842,"libraries":["datasets","mlcroissant"],"formats":[],"modalities":["video"]},"private":false,"repoType":"dataset","likes":0,"isLikedByUser":false,"isBenchmark":false},{"author":"MJ-Bench","downloads":18,"gated":false,"id":"MJ-Bench/MJ-Bench-Results","lastModified":"2024-07-09T07:23:39.000Z","datasetsServerInfo":{"viewer":"preview","numRows":0,"libraries":[],"formats":[],"modalities":[]},"private":false,"repoType":"dataset","likes":0,"isLikedByUser":false,"isBenchmark":false}],"models":[{"author":"MJ-Bench","authorData":{"_id":"666bb425066ceea60c0d59b8","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64d448d3d095077728f7c740/eiPzzUiEPaUmUeJRB7oRy.jpeg","fullname":"MJ-Bench-Team","name":"MJ-Bench","type":"org","isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":12,"isUserFollowing":false},"downloads":1,"gated":false,"id":"MJ-Bench/MJ-VIDEO-2B","availableInferenceProviders":[],"lastModified":"2025-02-06T05:13:19.000Z","likes":0,"private":false,"repoType":"model","isLikedByUser":false,"numParameters":2214241075},{"author":"MJ-Bench","authorData":{"_id":"666bb425066ceea60c0d59b8","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64d448d3d095077728f7c740/eiPzzUiEPaUmUeJRB7oRy.jpeg","fullname":"MJ-Bench-Team","name":"MJ-Bench","type":"org","isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":12,"isUserFollowing":false},"downloads":0,"gated":false,"id":"MJ-Bench/DDPO-alignment-gpt-4v","availableInferenceProviders":[],"lastModified":"2024-07-09T07:25:19.000Z","likes":0,"pipeline_tag":"text-to-image","private":false,"repoType":"model","isLikedByUser":false},{"author":"MJ-Bench","authorData":{"_id":"666bb425066ceea60c0d59b8","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64d448d3d095077728f7c740/eiPzzUiEPaUmUeJRB7oRy.jpeg","fullname":"MJ-Bench-Team","name":"MJ-Bench","type":"org","isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":12,"isUserFollowing":false},"downloads":0,"gated":false,"id":"MJ-Bench/DDPO-alignment-gpt-4o","availableInferenceProviders":[],"lastModified":"2024-07-09T07:25:18.000Z","likes":0,"pipeline_tag":"text-to-image","private":false,"repoType":"model","isLikedByUser":false},{"author":"MJ-Bench","authorData":{"_id":"666bb425066ceea60c0d59b8","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64d448d3d095077728f7c740/eiPzzUiEPaUmUeJRB7oRy.jpeg","fullname":"MJ-Bench-Team","name":"MJ-Bench","type":"org","isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":12,"isUserFollowing":false},"downloads":0,"gated":false,"id":"MJ-Bench/DDPO-alignment-claude3-opus","availableInferenceProviders":[],"lastModified":"2024-07-09T07:25:16.000Z","likes":0,"pipeline_tag":"text-to-image","private":false,"repoType":"model","isLikedByUser":false},{"author":"MJ-Bench","authorData":{"_id":"666bb425066ceea60c0d59b8","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64d448d3d095077728f7c740/eiPzzUiEPaUmUeJRB7oRy.jpeg","fullname":"MJ-Bench-Team","name":"MJ-Bench","type":"org","isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":12,"isUserFollowing":false},"downloads":5,"gated":false,"id":"MJ-Bench/DiffusionDPO-alignment-claude3-opus","availableInferenceProviders":[],"lastModified":"2024-07-09T07:22:28.000Z","likes":0,"pipeline_tag":"text-to-image","private":false,"repoType":"model","isLikedByUser":false},{"author":"MJ-Bench","authorData":{"_id":"666bb425066ceea60c0d59b8","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64d448d3d095077728f7c740/eiPzzUiEPaUmUeJRB7oRy.jpeg","fullname":"MJ-Bench-Team","name":"MJ-Bench","type":"org","isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":12,"isUserFollowing":false},"downloads":3,"gated":false,"id":"MJ-Bench/DiffusionDPO-alignment-gemini-1.5","availableInferenceProviders":[],"lastModified":"2024-07-09T07:21:56.000Z","likes":0,"pipeline_tag":"text-to-image","private":false,"repoType":"model","isLikedByUser":false}],"paperPreviews":[],"spaces":[{"author":"MJ-Bench","authorData":{"_id":"666bb425066ceea60c0d59b8","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64d448d3d095077728f7c740/eiPzzUiEPaUmUeJRB7oRy.jpeg","fullname":"MJ-Bench-Team","name":"MJ-Bench","type":"org","isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":12,"isUserFollowing":false},"colorFrom":"green","colorTo":"indigo","createdAt":"2024-06-30T07:02:38.000Z","emoji":"🥇","id":"MJ-Bench/MJ-Bench-Leaderboard","lastModified":"2024-07-13T14:35:08.000Z","likes":10,"pinned":true,"private":false,"sdk":"gradio","repoType":"space","runtime":{"stage":"RUNTIME_ERROR","hardware":{"current":null,"requested":"cpu-basic"},"storage":null,"gcTimeout":172800,"errorMessage":"Workload evicted, storage limit exceeded (50G)","replicas":{"requested":1},"devMode":false,"domains":[{"domain":"mj-bench-mj-bench-leaderboard.hf.space","stage":"READY"}]},"title":"MJ Bench Leaderboard","isLikedByUser":false,"originRepo":{"name":"demo-leaderboard-backend/leaderboard","author":{"_id":"655dbd8360009b03e4451217","avatarUrl":"https://www.gravatar.com/avatar/48236a8e5b71950f0708b3f2e3e7925f?d=retro&size=100","fullname":"Demo leaderboard with an integrated backend","name":"demo-leaderboard-backend","type":"org","isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":22,"isUserFollowing":false}},"ai_short_description":"Display and filter multimodal model leaderboard results","ai_category":"Data Visualization","trendingScore":0,"tags":["gradio","region:us"],"featured":false}],"buckets":[],"numBuckets":0,"numDatasets":3,"numModels":6,"numSpaces":2,"lastOrgActivities":[{"time":"2026-01-12T10:37:28.991Z","user":"zhuokai","userAvatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/63af25605fe9db73f67a0fb7/mcQHbTdJVxi2GdHzRB9ad.jpeg","type":"paper","paper":{"id":"2410.12138","title":"Preference Optimization with Multi-Sample Comparisons","publishedAt":"2024-10-16T00:59:19.000Z","upvotes":0,"isUpvotedByUser":false}},{"time":"2026-01-12T10:36:19.633Z","user":"zhuokai","userAvatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/63af25605fe9db73f67a0fb7/mcQHbTdJVxi2GdHzRB9ad.jpeg","type":"paper","paper":{"id":"2601.05106","title":"Token-Level LLM Collaboration via FusionRoute","publishedAt":"2026-01-08T16:53:16.000Z","upvotes":40,"isUpvotedByUser":true}},{"time":"2025-11-10T09:32:09.130Z","user":"zhuokai","userAvatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/63af25605fe9db73f67a0fb7/mcQHbTdJVxi2GdHzRB9ad.jpeg","type":"paper","paper":{"id":"2511.03773","title":"Scaling Agent Learning via Experience Synthesis","publishedAt":"2025-11-05T18:58:48.000Z","upvotes":82,"isUpvotedByUser":true}}],"acceptLanguages":["*"],"canReadRepos":false,"canReadSpaces":false,"blogPosts":[],"currentRepoPage":0,"filters":{},"paperView":false}">