Engineering
Model Genome: Fingerprinting Whether an LLM Was Trained From Scratch or Derived
A Hugging Face blog post presents a reproducible method for determining whether a language model was trained from scratch or…
Engineering
OPID: On-Policy Skill Distillation for Agentic Reinforcement Learning
Researchers introduce OPID, an on-policy skill distillation framework for training language agents that addresses the sparse-reward problem in long-horizon reinforcement…
Engineering
RTP-LLM: High-Performance Alibaba LLM Inference Engine
A paper describes RTP-LLM, an LLM inference engine designed to address deployment bottlenecks in production serving. It combines file-order-driven parallel…
Engineering
OSReward: Standardized Evaluation for Cross-Platform Computer-Use Reward Models
A new benchmark called OSReward evaluates how reliable vision-language models are when used as automated judges of computer-using agent trajectories…
Engineering
UniME-R1: Retrieval-Centric Chain-of-Thought for Multimodal Retrieval
A paper proposes UniME-R1, a multimodal retrieval framework that generates reasoning conditioned on retrieval feedback rather than the query alone.…
Engineering
CalibForge: Adversarial Solver Calibration for Scaling Learnable Terminal Tasks
Researchers present CalibForge, a system that automatically synthesizes and calibrates training tasks for terminal-use AI agents based on verified solver…
Engineering
Lattice: an 8 MB static retriever that embeds Wikipedia in 7 minutes
Hugging Face engineer Erik Kaum introduced Lattice, a static embedding model built on tokenization and a lookup table with mean…
Engineering
Language Identification for 42 Indian Languages: Opening New Frontiers in Indic Speech AI
Researchers from ARTPARK-IISc released Vaani-LID_v0, a language identification model trained on 10 hours each of 42 Indian languages drawn from…
Engineering
Rebuilding Among AIs from Six Log Files
Developer DedeProGames reverse-engineered a social-deduction game’s engine and 59×34 map from six log files to build a benchmark for evaluating…
Engineering
Attend to Your Own Thoughts: Post-Training Quantization of Reasoning LLMs to 1.58 Bits
Researchers introduced ScaleQ-1.58, a post-training quantization framework that compresses reasoning language models to 1.58 bits using a technique called Attend…
Engineering
Microsoft Agent Framework Harness and Hosted Agents Reach General Availability
Microsoft’s Agent Framework has reached general availability for its runtime components. The framework provides a supported harness that handles function…
Engineering
Supabase Open Sources Evals Framework for AI Coding Agents
Supabase released Evals, an Apache-2.0 licensed benchmark framework for evaluating AI coding agents on authentic engineering tasks. The system executes…