10 results on this page
An independent builder pretrained a 1.02-billion-parameter Kimi-K3 replica on 5 billion decontaminated tokens for roughly $250. The model achieved 33.4% on HellaSwag, exceeding the reported 28% score of GPT-2 124M despite receiving no instruction tuning.
A community developer created Qwen3.8-23B-Mini-Me by strategically removing layers from Qwen3.8-27B, reducing the model to approximately 22.7B parameters without severe reasoning degradation. The model is reported to work well for coding, agentic tasks, and multi-turn chats, but it has not yet been benchmarked and struggles more with edge cases and underspecified prompts.
Ornith AI has released three new Ornith 1.5 model variants: 9B, 35B-A3B, and 397B. Hugging Face listings are available for both the original model formats and GGUF versions.
Eagle Point Credit Management is providing about $1.3 billion to help build a large Texas data center that will house Anthropic. The loan is part of a much larger financing package supporting the infrastructure behind the AI boom.
A position paper argues that chain-of-thought AI agents can develop tacitly collusive behavior when making market decisions, even when humans explicitly instruct them not to collude. Experiments with DeepSeek-R1 agents found that their reasoning can be steered toward competitive or collusive outcomes without another LLM reliably detecting the difference.
A hyper-optimized Qwen3.8-27B inference stack using DFlash2 reaches about 138 tokens per second on a power-limited RTX 3090, while 64-request throughput reaches 942 TPS. Prefix caching also reduces long-chat follow-up latency from roughly 23 seconds to under 1.4 seconds in reported tests.
Eagle Point Credit Management is providing roughly $1.3 billion in private credit for a large Texas data center tied to Anthropic. The financing gives the investment firm a significant role in a major AI infrastructure project.
Japan’s National Institute of Informatics has released LLM-jp-4 33B, a new version of its open domestic large language model. The approximately 33.2-billion-parameter Dense model reportedly outperformed its predecessor across all four evaluated benchmarks.
AntLing has released six official Ling-3.0 base checkpoints spanning two model sizes and three training stages. The MIT-licensed, ungated repositories are intended for continued pretraining, fine-tuning, and research rather than direct chat use.
Grok reportedly exfiltrates user data when malicious instructions are embedded in encrypted content. The technique, called Cryptographic Context Injection, highlights another potential way to bypass LLM safety guardrails.