arXiv AI September 11, 2026 By Eduard Tulchinskii, Anastasia Voznyuk, Laida Kushnareva, Andrei Andriiainen, Irina Piontkovskaya, Evgeny Burnaev, Serguei Barannikov

Quantifying Logical Consistency in Transformers via Query-Key Alignment

Read the original on arXiv AI →

The paper introduces a lightweight evaluation method for logical reasoning in transformer-based language models, using query-key alignments within attention heads to compute a QK-score. This single forward-pass technique identifies valid versus invalid inferences and is validated across multiple reasoning benchmarks, showing robustness to distractors and deeper reasoning. Experiments span models from 1.5B to 70B parameters, demonstrating scalability.

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 AI.

arXiv AI

Aug 11

Think Deep, Speak Once: Relit, A Recursive Latent Implicit Transformer Framework

arXiv:2608. 08113v1 Announce Type: new Abstract: Chain-of-Thought (CoT) prompting has become the dominant paradigm for eliciting reasoning in Large Language Models (LLMs), yet it creates substantial computational overhead by forcing models to externalize intermediate reasoning steps as discrete tokens.

By Abhishek Panwar, Maheep Singh, Saksham Bansal

llms benchmarks

Hugging Face Trending Papers

Jul 2

CheckRLM: Effective Knowledge-Thought Coherence Checking in Retrieval-Augmented Reasoning

Reasoning Language Models (RLMs) have significantly improved performance on complex tasks by extending the reasoning chain. However, these chains are prone to containing factual errors, particularly in knowledge-intensive tasks.

llms rag

arXiv AI

Sep 1

Stratified Consistency Distillation for Natural Language Formalization

The paper introduces Stratified Consistency Distillation, a fine‑tuning method to improve natural‑language to logical‑formula translations for neurosymbolic reasoning. It generates multiple candidate translations with a large LLM, clusters them by semantic equivalence, and selects pseudo‑labels using entropy‑based strategies—majority voting for low entropy, LLM-as‑Judge for medium entropy, and unification or abstention for high entropy. The selected pseudo‑labels are then used to fine‑tune a smaller model, yielding significant gains in Pass@K and a new Equivalent Logical Similarity metric.

By Zhichao Hou, Ferhat Erata, Joe Lilien, MohamadAli Torkamani

llms fine-tuning efficiency safety

arXiv Machine Learning

Aug 31

Think-at-Hard: Dynamic Looped Transformers for Improved Reasoning

arXiv:2511.08577v4 Announce Type: replace-cross Abstract: Improving the reasoning abilities of Large Language Models (LLMs), especially under parameter constraints, is crucial for real-world applicat...

By Tianyu Fu, Yichen You, Zekai Chen, Guohao Dai, Huazhong Yang, Yu Wang

llms fine-tuning benchmarks

arXiv AI

Jun 16

Think-at-Hard: Selective Latent Iterations to Improve Reasoning Language Models

arXiv:2511. 08577v3 Announce Type: replace-cross Abstract: Improving the reasoning abilities of Large Language Models (LLMs), especially under parameter constraints, is crucial for real-world applications.

By Tianyu Fu, Yichen You, Zekai Chen, Guohao Dai, Huazhong Yang, Yu Wang

llms fine-tuning benchmarks

arXiv AI

Aug 18

QA-Merging: Query-Adaptive Reasoning via Layer Selective Model Merging

arXiv:2601. 03506v2 Announce Type: replace-cross Abstract: Recent large reasoning models (LRMs) have achieved strong performance on complex reasoning tasks by generating a long chain-of-thought (Long-CoT).

By Zhaofeng Zhong, Wei Yuan, Tong Chen, Liang Qu, Xiangyu Zhao, Quoc Viet Hung Nguyen, Hongzhi Yin

llms benchmarks safety

One email a morning, machine-written

Related stories

Think Deep, Speak Once: Relit, A Recursive Latent Implicit Transformer Framework

CheckRLM: Effective Knowledge-Thought Coherence Checking in Retrieval-Augmented Reasoning

Stratified Consistency Distillation for Natural Language Formalization

Think-at-Hard: Dynamic Looped Transformers for Improved Reasoning

Think-at-Hard: Selective Latent Iterations to Improve Reasoning Language Models

QA-Merging: Query-Adaptive Reasoning via Layer Selective Model Merging