arXiv Machine Learning July 31, 2026 By Lei Dong

The Kinetics of Training: A Driven-Nucleation Rate Law for Emergence, Plasticity Loss, and Circuit Control in Language Models

Read the original on arXiv Machine Learning →

arXiv:2607. 27281v1 Announce Type: new Abstract: A capability appears in a language model when the last parts of its circuit align in one stochastic attempt, and getting all but one right is worth nothing.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Machine Learning.

arXiv Machine Learning

Sep 1

Ceiling-Clipped Acceptance Histograms Indicate Stranded Speed-up in Block-Diffusion Speculative Decoding

arXiv:2608.30427v1 Announce Type: cross Abstract: Speculative decoding speeds up generation with an efficient draft model (drafter) that proposes tokens for a target model to verify in one pass, pres...

By Ephrem Wu

diffusion fine-tuning benchmarks

arXiv AI

Jul 22

Operational Proto-Introspection in Looped Language Models: Process-Quality Taps, Executable Branching, and the Readout-Control Boundary

arXiv:2607. 18553v1 Announce Type: cross Abstract: Can a language model read the quality of ongoing computation, and can an external intervention turn that readout into better outcomes?

By Jan Kirin

llms fine-tuning

arXiv AI

Jul 22

Circuit Claims Depend on What Is Extracted and How It Is Compared

arXiv:2607. 18921v1 Announce Type: cross Abstract: Circuit extraction identifies a small set of model components whose presence preserves a target behavior under ablation, and the resulting circuit is often read as the mechanism behind that behavior.

By Yang Sheng, Jie Fu

llms reinforcement-learning efficiency benchmarks

arXiv AI

Sep 10

The Geometry of Refusal: Why Post-Hoc Safety Is Fragile and Pretraining-Time Safety Persists

arXiv:2609.06934v1 Announce Type: cross Abstract: Post-hoc safety training (RLHF, DPO) is the dominant way to align large language models, yet jailbreaks (Zou et al., 2023b), fine-tuning attacks (Qi...

By Srikanth Malla, Chiho Choi, Joon Hee Choi

llms reinforcement-learning fine-tuning safety

arXiv AI

Jun 2

When Do Attention Circuits Form? Developmental Trajectories of Capability and Attention-Sink Emergence Across Three 1B-ClassArchitectures

arXiv:2606. 02378v1 Announce Type: cross Abstract: We track the developmental trajectory of attention-head circuit formation across three 1B-class language models spanning two architecture families (dense transformer, mixture-of-experts) and two pretraining corpora (The Pile, DCLM): Pythia 1B, OLMo 1B-0724-hf, and OLMoE 1B-7B-0924.

By Yongzhong Xu

llms safety

arXiv AI

Aug 25

Crossing the Margin Cliff: Toward Relearn-Robust LLM Unlearning via Margin Calibration

arXiv:2607.27836v2 Announce Type: replace Abstract: Large language model unlearning is consistently fragile under relearn attacks. On TOFU, fine-tuning on twenty forget examples substantially recover...

By Xiangyu Yin, Jiaxu Liu, Zhen Chen, Chih-Hong Cheng

llms fine-tuning efficiency

One email a morning, machine-written

Related stories

Ceiling-Clipped Acceptance Histograms Indicate Stranded Speed-up in Block-Diffusion Speculative Decoding

Operational Proto-Introspection in Looped Language Models: Process-Quality Taps, Executable Branching, and the Readout-Control Boundary

Circuit Claims Depend on What Is Extracted and How It Is Compared

The Geometry of Refusal: Why Post-Hoc Safety Is Fragile and Pretraining-Time Safety Persists

When Do Attention Circuits Form? Developmental Trajectories of Capability and Attention-Sink Emergence Across Three 1B-ClassArchitectures

Crossing the Margin Cliff: Toward Relearn-Robust LLM Unlearning via Margin Calibration