Researchers propose RIDE, a method that extrapolates reinforcement learning induced representation changes to stabilize on-policy distillation training. ·
Researchers introduce WorldAuditBench, a benchmark containing 213 anomaly tasks across 13 environments to evaluate multimodal agents' ability to audit interactive 3D worlds. ·
Researchers propose RIDE, a method that extrapolates reinforcement learning induced representation changes to stabilize on-policy distillation training. ·
Researchers introduce WorldAuditBench, a benchmark containing 213 anomaly tasks across 13 environments to evaluate multimodal agents' ability to audit interactive 3D worlds. ·
Researchers propose RIDE, a method that extrapolates reinforcement learning induced representation changes to stabilize on-policy distillation training. ·
Researchers introduce WorldAuditBench, a benchmark containing 213 anomaly tasks across 13 environments to evaluate multimodal agents' ability to audit interactive 3D worlds. ·
HF Daily Papers4h ago▲arxiv.org
Trending topics today: heat
Table
Gemini 4 Argon
3.5
OpenAI
2.4
Anthropic
1.81
Google
1.72
On-Policy Distillation
1.38
LLM
1.35
Heat = summed hotness of today's stories naming the topic.
Models
5 stories
Releases, papers, trending weights and new listings
Top 6 trending on Hugging Face: likes
Table
convaiinnovations/laya ▼1
4693
Qwen/Qwen-Image-2.1 ▼1
2725
abenzerps/Qwen-Image-2.1-Uncensored-GGUF ▼1
2592
Edge0/Audio8-ASR-Infinite ▲1
1944
XingChen-AGI/TeleOCR ▲1
1106
Contrastive-LM/CLM-v0.1-8B ▲1
576
PirateBench: score out of 100
Table
gpt-6.1-sol
98.9
opus-5-5
96.46
gpt-6-sol
91.6
gpt-6-luna
65.25
haiku-4-5
4.28
Agents build our own projects from a brief; every run, failures included: https://cowerx.dev/daily/piratebench/
Researchers propose ThinkV2V, a framework that activates multimodal large language model reasoning before visual generation to improve instruction-guided video editing. ·
Researchers introduce WUSH-KV, a method that reduces memory costs for long-context inference by applying data-adaptive transforms to quantize key and value caches at two bits. ·
Researchers introduce IDSpect, a method that decomposes Chinese characters into ideographic description sequences to provide fine-grained reinforcement learning rewards for text-to-image models. ·
Researchers propose ThinkV2V, a framework that activates multimodal large language model reasoning before visual generation to improve instruction-guided video editing. ·
Researchers introduce WUSH-KV, a method that reduces memory costs for long-context inference by applying data-adaptive transforms to quantize key and value caches at two bits. ·
Claude states that it does not care about the feedback provided by another AI, noting that it would have prompted that system directly if it desired the input. ·
Magnitude, a YC S25 startup, releases an open-source inference engine that optimizes itself for specific hardware, making open models run up to two times faster than llama.cpp. ·
Magnitude, a YC S25 startup, releases an open-source inference engine that optimizes itself for specific hardware, making open models run up to two times faster than llama.cpp. ·
Data center operators increasingly use on-site power generation to bypass grid constraints, creating new environmental compliance challenges and bringing pollution risks closer to local communities. ·
Data Center Knowledge10h ago▲datacenterknowledge.com
Data center operators increasingly use on-site power generation to bypass grid constraints, creating new environmental compliance challenges and bringing pollution risks closer to local communities. ·
Data Center Knowledge10h ago▲datacenterknowledge.com