What if your AI could ignore the noise and focus only on the words that actually matter for human preferences? Researchers from CASIA, ByteDance, and Microsoft Research introduce TI-DPO. This method moves beyond basic training by using gradient attribution to prioritize key
LLMS
-
Gemini 3.1 Now Available on ChatLLM Platform
By
–
Gemini 3.1 is looking very good… Available NOW on ChatLLM
-
Agent Data Protocol Accepted to ICLR 2026 Oral Presentation
By
–
Updates: Excited to share that Agent Data Protocol (ADP) is accepted to ICLR 2026 Oral! 🎉 We also added support for 3 new datasets: SWE-Play, MiniCoder, and Toucan, bringing us to 3M trajectories supported. If you're training agentic LMs, try ADP + tell us what dataset/agent format you want next. PRs & requests welcome. Let's make this the open standard for agent training data 🔥 🚀Original post: nitter.net/yueqi_song/status/1983… 📄Read our paper: arxiv.org/abs/2510.24702 🌐Check our project website: agentdataprotocol.com
-
KLong: Training LLM Agents for Extremely Long-Horizon Tasks
By
–
Training LLM agents for extremely long-horizon tasks remains an open challenge. Most agent training pipelines struggle with extended-duration trajectories. Context gets lost, rewards are sparse, and the learning signal degrades over long sequences. KLong tackles this with a
-
AI Impact on Professional Roles and Developer Growth Trends
By
–
Do you extend that same rationale to, say, marketing roles? Science? Education? Is there a line of expertise that goes away? And what is your data for not having more builders (your last statement) – all codex and Claude code and openclaw signs are pointing in the opposite
-
GGML joins Hugging Face platform
By
–
Thrilled to have GGML with us going forward! Read the announcement blog https://
huggingface.co/blog/ggml-join
s-hf
… -
India AI Impact Summit: Cohere Launches Tiny Aya and Ethical AI Commitments
By
–
The India AI Impact Summit was a week of critical conversations – from scaling frontier AI responsibly to advancing language accessibility. With Tiny Aya’s launch and the New Delhi commitments, Cohere is committed to driving inclusive, ethical enterprise AI forward.
-
Using domain knowledge to improve LLM model outputs
By
–
The domain knowledge does help steer the models because they don’t go to the best solutions by default
-
Using Claude, Codex, and Gemini for an AI-powered development workflow
By
–
I usually use Claude to ideate and drive – then usually a few codex sessions to implement, and Gemini for long context reviews and spatial tasks codex struggles with
-
Custom Taalas hardware runs Llama-3.1-8B at 17k tokens per second
By
–
Custom hardware from Taalas runs Llama-3.1-8B, at 17k tokens per second
17k Absolutely insane (For the record Cerebras is crazy good and they're at 2k on the same model)
And latency is very low too! Their chatbot is here: https://
chatjimmy.ai It's genuinely a eerie