Iris-3B Preprint Explores Pixel-Space Diffusion Scaling and Downstream Tasks
The October 2026 arXiv preprint presents Iris-3B, a 3B-parameter pixel-space text-to-image transformer pretrained from scratch, and reports negative results when testing…
More freedom.
More control.
New AI releases, creative tools and what you can actually download, change and use.
118 stories
The October 2026 arXiv preprint presents Iris-3B, a 3B-parameter pixel-space text-to-image transformer pretrained from scratch, and reports negative results when testing…
arXiv preprint benchmarks three frontier models on entity alignment tasks across three datasets and a new biomedical benchmark, revealing that label exposure causes judges to…
An arXiv preprint dated 8 October 2026 introduces KPSA, a kernel-based method derived from the empirical neural tangent kernel that approximates input-dependent parameter…
An arXiv preprint released October 9 2026 finds attention heads, MLP neurons and sparse autoencoder features whose activations correlate with narrativity in GPT-2-Small…
The October 2026 CliQ-RRG preprint outlines a clinical-knowledge guided approach using disease-aware visual-textual alignment and graph retrieval to produce concise QA-format…
arXiv preprint presents DualQuant, a calibration-free post-training method that minimizes weight-reconstruction MSE and yields per-channel activation rescalings as a byproduct…
arXiv preprint presents TQS, a linear-time framework that scores coding-agent trajectories by structural convergence signals instead of final benchmark outcomes, improving…
arXiv preprint dated 2026-10-09 introduces CAST, a shift-or-stay framework that redistributes chunk-level advantages using predictive uncertainty and teacher-signal smoothness…
arXiv preprint from 2026-10-07 shows a deployed fine-tuned 0.8B Qwen3.5 model beats prompted GPT-5.4 and GPT-5.6 in precision and recall for grammar mastery tracking from…
New preprint introduces FedAlphaEdit, a framework that combines null-space editing with server-side merging so multiple parties can jointly update an LLM without sharing raw…
Preprint introduces model-agnostic framework that incorporates external ranking priors via temperature-scaled KL divergence to regularize credit models when labeled data…
A new arXiv preprint shows that a restricted form of RAM-Net can theoretically bridge per-KV discrete compression and recurrent multi-KV aggregation, enabling weight migration…