
InternLM's JanusCoder Beats GPT-4o at Turning Screenshots Into Working Code
InternLM's new open model suite unifies chart-to-code, web UI generation, and Manim animation into one visual-programmatic interface
20 stories

InternLM's new open model suite unifies chart-to-code, web UI generation, and Manim animation into one visual-programmatic interface

A dual-brain robot framework pairs a slow planner with a fast video watcher, cutting task time and boosting success rates on messy real-world jobs.

A research team unveils an interactive Europe-wide biodiversity mapping platform powered by CNN ensembles and a species-focused BERT model at 50-meter resolution.

A new vector lookup table design cuts parallel LLM inference bottlenecks on CPUs, delivering up to 4.2x speedup for ternary models on edge devices.

A training-free method probes multimodal models with paraphrased prompts to catch and correct hallucinations driven by language priors, without touching the image or the weights.

A frequency-domain neural operator uses Euler's formula to split satellite image fusion into phase and magnitude, hitting top accuracy with 264K parameters.

A new black-box attack reconstructs the hidden agent graph inside LLM multi-agent systems with AUC up to 0.99, exposing IP and attack surfaces.

A new filtering strategy for explainable clustering cuts redundant symbolic patterns, slashing ILP solve times by up to 5x while preserving or improving cluster quality.

A new framework automatically picks the best counting line in traffic video, hitting near-perfect accuracy while running up to four times faster than full-frame processing.

A neurosymbolic pipeline uses PPO and a graph neural network to fix planner trajectories that ignore acceleration limits and drag.

A signal processing approach uses tunable beta-divergence factorization to pull out overlapping voices from reverberant multichannel recordings, beating two blind separation baselines.

A Gabor dictionary paired with elastic net regularization feeds a 1D+2D CNN-LSTM classifier, hitting 98.95% on five heart valve conditions.

Researchers show that soft robots trained with knowledge from peers learn faster than isolated learners, and learning from multiple teachers yields more consistent gains.

A German research team built a reproducible pipeline that turns S2ORC into 582,683 chemistry papers with paragraph embeddings and audited open-access licenses.

New mechanistic interpretability work on Qwen-3 shows LLMs form internal representations of future words, and this planning ability scales with model size.

A new benchmark asks multimodal models to pick the sharper photo from expert-grade image pairs and explain their reasoning like a pro.

A four-level framework compares first-order logic arguments using SBERT embeddings, context weights, and symbolic structure, with axiomatic guarantees baked in.

A new SVD-guided training constraint keeps LoRA adapters from colliding across dozens of unlearning requests, avoiding costly router overhead at inference.

A pre-trained foundation model reads directly from multi-table SQL databases and beats supervised graph models on 41 predictive tasks without task-specific training.

Automated prompt evolution boosts LLM password cracking rates from 2.02% to 8.48%, exposing how small text tweaks can quadruple attacker success.