#language-models

The latest AI and developer news about #language-models, with the original source and a short summary.

Researchers extend prior-fitted networks to natural language with synthetic non-linguistic training

A 300M-parameter transformer trained on synthetic recurrent sequences achieves 0.9–2.4 bits per byte on real languages after reading 1 million bytes.

·2026.10.06 19:00

ASRN: Adaptive Sparse Recurrence Network introduces linear-memory copy layer for language models

The new ASRN architecture uses learned hash tables to locate and copy prior context occurrences, achieving memory usage linear in sequence length.

·1·2026.10.05 07:00

Apple Research Improves Discrete Flow Matching Trajectories with Energy Compass… 32% Perplexity Reduction vs. Teacher Model with 8-Step Inference

A study featured on the Apple Research blog proposes TS-DFM to improve the trajectory quality of discrete flow matching, reducing perplexity by 32% compared to a 1,024-step teacher model with 8-step inference and improving speed by 128x.

·1·2026.09.16 09:00

#language-models

Researchers extend prior-fitted networks to natural language with synthetic non-linguistic training

ASRN: Adaptive Sparse Recurrence Network introduces linear-memory copy layer for language models

Apple Research Improves Discrete Flow Matching Trajectories with Energy Compass… 32% Perplexity Reduction vs. Teacher Model with 8-Step Inference

talkie releases a language model from the 1930s

Entropy-Preserving Reinforcement Learning

GSPO for Scalable Reinforcement Learning of Language Models

Researchers extend prior-fitted networks to natural language with synthetic non-linguistic training

ASRN: Adaptive Sparse Recurrence Network introduces linear-memory copy layer for language models

Apple Research Improves Discrete Flow Matching Trajectories with Energy Compass… 32% Perplexity Reduction vs. Teacher Model with 8-Step Inference

talkie releases a language model from the 1930s

Entropy-Preserving Reinforcement Learning

GSPO for Scalable Reinforcement Learning of Language Models