arXiv Machine Learning June 25, 2026 By Randhir Kumar

Holographic Memory for Zero-Shot Compositional Reasoning in Knowledge Graphs: A Mechanistic Study of Where and Why It Fails

Read the original on arXiv Machine Learning →

arXiv:2606. 24948v1 Announce Type: new Abstract: Knowledge graph embedding (KGE) models predict single-hop links well but have no mechanism for zero-shot compositional queries: multi-hop questions whose relation chains never appeared during training.

Summary generated by The Flow from the publisher's feed. The full article lives at arXiv Machine Learning.

arXiv AI

Jul 7

Input Pathways Shape Few-Shot, Not Zero-Shot, Binding in Tiny Transformers: A Fully-Enumerable Study

arXiv:2607. 04926v1 Announce Type: cross Abstract: How does the way information reaches a transformer -- as symbolic tokens, a clean per-factor "oracle" code, or an entangled perceptual vector -- shape whether it binds that information compositionally?

By Yoshiyuki Ootani

llms safety

arXiv AI

Jun 26

Retrieval-Warmed Energy-Based Reasoning: A Five-Arm Ablation Methodology for Diffusion-as-Inference on Structured Reasoning Tasks

arXiv:2606. 26476v1 Announce Type: cross Abstract: Warm-started diffusion samplers accelerate iterative inference, but it is rarely clear which part of the pipeline carries the gain.

By Libo Sun, Po-Wei Harn, Zewei Zhang, Peixiong He, Xiao Qin

llms rag diffusion benchmarks safety

arXiv AI

Jul 2

DiscoLoop: Looping Discrete Embeddings and Continuous Hidden States for Multi-hop Reasoning

arXiv:2607. 00341v1 Announce Type: cross Abstract: Large language models achieve strong performance on many reasoning tasks when allowed to externalize intermediate steps as Chain-of-Thought (CoT).

By Hengyu Fu, Tianyu Guo, Zixuan Wang, Hanlin Zhu, Jason D. Lee, Jiantao Jiao, Stuart Russell, Song Mei

llms rag benchmarks

arXiv Machine Learning

2d ago

You Only Pass Once: Answering and Abstaining Together in a Single Forward Pass of a Frozen Language Model

arXiv:2608. 14465v1 Announce Type: cross Abstract: A frozen language model on reasoning tasks has two coupled weaknesses: it under-uses evidence its own residual stream already encodes, and it fails to detect when the input is insufficient to answer, so it confabulates.

By Ziyang Luo, Zhongyao Chu, Xinjie He, Youting Wang, Xukui Qin, Runxiong Wu, Yan-Syuan Chen

llms benchmarks

arXiv Machine Learning

Jul 1

Explicit Fuzzy Logic in the Feed-Forward Layer: Self-Forgetting Quantifiers Discover Legible Grammatical-Licensing Detectors

arXiv:2606. 31845v1 Announce Type: cross Abstract: A transformer's feed-forward (FFN) sublayer materializes the distinctions attention gathers, yet gives no account of what it computes.

By Mark Oskin

llms

arXiv AI

Jul 23

Profile-Graph Memory for LLM Agents: Implicit Cross-Entity Traversal through Narrative Profiles

arXiv:2607. 19359v1 Announce Type: new Abstract: Long-term memory is essential for LLM agents that interact across sessions, yet current memory benchmarks primarily evaluate single-hop recall, leaving multi-hop association largely unmeasured.

By Shengtong Zhu

llms rag agents benchmarks

Related stories

Input Pathways Shape Few-Shot, Not Zero-Shot, Binding in Tiny Transformers: A Fully-Enumerable Study

Retrieval-Warmed Energy-Based Reasoning: A Five-Arm Ablation Methodology for Diffusion-as-Inference on Structured Reasoning Tasks

DiscoLoop: Looping Discrete Embeddings and Continuous Hidden States for Multi-hop Reasoning

You Only Pass Once: Answering and Abstaining Together in a Single Forward Pass of a Frozen Language Model

Explicit Fuzzy Logic in the Feed-Forward Layer: Self-Forgetting Quantifiers Discover Legible Grammatical-Licensing Detectors

Profile-Graph Memory for LLM Agents: Implicit Cross-Entity Traversal through Narrative Profiles