New research probes LLM reasoning, concept discovery, and knowledge handling
ByPulseAugur Editorial·[20 sources]·
Recent research explores the internal workings and reasoning capabilities of large language models (LLMs). Studies are investigating how LLMs process information beyond linear assumptions, examining the coupling between internal and verbalized probabilities, and developing frameworks to interpret reasoning trajectories. Additionally, researchers are evaluating LLMs' ability to handle multi-valued relations, their selective reliance on external guidance, and their understanding of ownership concepts.
AI
IMPACT
These studies offer deeper insights into LLM interpretability, reasoning, and knowledge handling, potentially guiding future model development and application.
RANK_REASON
Cluster consists of multiple academic papers published on arXiv.
arXiv:2610.09569v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly used for emotional support, raising concern that sustained use may draw users away from their real-world relationships. Yet existing evaluations primarily focus on the safety, empathy, o…
arXiv cs.AI
TIER_1English(EN)·Everett Rush, David J. Icove, Ari Kim, Byung H. Park, Michael A. Langston·
arXiv:2610.02254v1 Announce Type: cross Abstract: Interpretive Structural Modeling (ISM) is a well-known process for multi-criteria decision making. The success of ISM over other methodologies is its ability to model causal relationships, the binary scale of factors, and resultin…
arXiv:2610.03052v1 Announce Type: new Abstract: Large language models (LLMs) contain broad knowledge, but they cannot access all of it reliably. We study this problem through knowledge accessibility, which describes whether the knowledge needed for a query can be recalled from th…
arXiv cs.LG
TIER_1English(EN)·Tido Specht, Elias Benedict Krey, Nils Neukirch, Nils Strodthoff·
arXiv:2610.01821v1 Announce Type: new Abstract: Understanding information processing in large language models (LLMs) requires dissecting the geometric organization of their internal token representations. While existing mechanistic interpretability (MI) methods seek to extract co…
arXiv:2610.00827v1 Announce Type: cross Abstract: Large language models carry an internal notion of uncertainty in their sampling distribution, i.e., the probabilities they place on generating one answer rather than another. They can also be asked to state a confidence, in words …
arXiv:2610.00571v1 Announce Type: cross Abstract: Large reasoning models (LRMs) have achieved substantial improvements in solving complex mathematical problems, but often produce lengthy, repetitive, or erroneous reasoning trajectories. In this work, we introduce a new interpreta…
arXiv:2609.39483v1 Announce Type: new Abstract: Ownership establishes rights over the use, control, and transfer of objects. Understanding these relations is essential for AI systems to interact appropriately with people and their resources. Yet how large language models (LLMs) a…
arXiv cs.CL
TIER_1English(EN)·Lukas Thede, Yash Kumar Atri, David Chen, Danielle Bitterman, Matthias Bethge, Tom Hartvigsen, Zeynep Akata·
arXiv:2609.38543v1 Announce Type: new Abstract: Constantly evolving real-world knowledge necessitates models to be updated continuously. Especially in medicine, as clinical evidence changes over time, outdated knowledge can pose safety risks. Existing evaluations of knowledge int…
arXiv:2609.39069v1 Announce Type: new Abstract: Test-time reasoning systems often respond to failure by restarting or revising the latest step, even when an earlier decision caused the error. We introduce CORE, a search controller that requests a certified conflict core from a ve…
arXiv:2609.39578v1 Announce Type: new Abstract: Agent harnesses often improve language models with human-designed workflows, but as models grow more capable, unreliable guidance can increasingly constrain their execution. We call the ability to benefit from useful guidance while …
arXiv cs.AI
TIER_1English(EN)·Nishchal Prasad, Eric Gaussier, Emilie Devijver, Alexander Obeid Guzman, Armen Aghasaryan, Gregor G\"ossler·
arXiv:2609.39406v1 Announce Type: new Abstract: Causal AI is a branch of Artificial Intelligence which helps understand and reason about cause and effect relationships, not just patterns or correlations. Causal discovery aims to infer elements of the underlying causal structure--…
arXiv cs.CL
TIER_1English(EN)·Timo Pierre Schrader, Annemarie Friedrich, Simon Razniewski, Lukas Lange·
arXiv:2609.36209v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly used as knowledge bases (KBs) due to the vast amount of knowledge they acquire during pre-training. While many works focus on extracting single relational triples, most real-world relati…
arXiv cs.AI
TIER_1English(EN)·Huaman Sun, Dingcheng Wang, Jason Hartline, Jessica Hullman·
arXiv:2609.38005v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly proposed as decision assistants who must reason probabilistically from available evidence under explicit decision costs. We propose a decision-theoretic framework that decomposes LLMs' d…
arXiv cs.LG
TIER_1English(EN)·Yundaichuan Zhan, Weishi Wang, Wenbiao Liu, Daniel Dahlmeier, Chengwei Qin, Juncheng Li, Fredrik D. Johansson, Zhongqi Yue·
arXiv:2609.36116v1 Announce Type: new Abstract: We study how to build more capable general-purpose agents by extending large language models (LLMs) with native typed decision-making. We introduce Dyad, an architecture that augments a pretrained LLM with an environment-conditioned…
Agent harnesses often improve language models with human-designed workflows, but as models grow more capable, unreliable guidance can increasingly constrain their execution. We call the ability to benefit from useful guidance while overriding unreliable guidance thinking outside …
Large language models (LLMs) are increasingly proposed as decision assistants who must reason probabilistically from available evidence under explicit decision costs. We propose a decision-theoretic framework that decomposes LLMs' decision loss into two components: forming accura…
We introduce Context Language Models (CLMs), language models that natively manage their own context. We implement this by treating the context as a file and allowing the model to make unrestricted updates to this file. This allows the model to learn what is most important to main…
arXiv cs.AI
TIER_1English(EN)·Xiangwei Wang, Wei Wang, Ken Chen, Nanduni Nimalsiri, Sachith Seneviratne, Saman Halgamuge·
arXiv:2602.01034v2 Announce Type: replace Abstract: Reinforcement learning (RL) has become a widely used paradigm for improving the reasoning abilities of large language models (LLMs) and Vision-language models (VLMs). Sparse binary outcome rewards, however, score only final corr…
Large language model (LLM) routing aims to select the most suitable model for each incoming query. Most existing routers learn this decision directly from query embeddings, model representations, preference data, or clusters of similar examples. Such approaches can be effective, …
arXiv:2610.02876v1 Announce Type: new Abstract: A multimodal large language model that correctly reports a spatial fact does not necessarily use that fact in subsequent reasoning. To study this distinction, we introduce \textsc{SpaceConflict}, a benchmark of 23{,}196 inputs for t…