arXiv subjects

Zifei Yu

Publications and source records attributed to Zifei Yu.

5 recordsLinked to original sources

arXiv2026

Understanding the Self-Reflection Mechanisms of LLMs through Biased Attitude Associations

While the emergent self-reflection capabilities of Large Language Models (LLMs) offer a promising paradigm for autonomous bias mitigation, their internal mechanics remain unclear, raising concerns regarding potential bias entrenchment. Under the premise that social bias is intrinsically encoded as valence inclinations, where the exacerbation of bias scales with sharper valence fluctuations across social groups, this paper proposes ReBias-Lens, a probing framework designed to interpret how self-reflection reconfigures these biased attitude associations through the lens of valence projection within intersectional contexts. Central to ReBias-Lens is the metric of Valence Fluctuation (VF) comprising two variants: Global-VF, which captures macroscopic valence encoding trends, and Local-VF, which scrutinizes microscopic distinctiveness across specific social categories. Deploying ReBias-Lens to evaluate four LLMs across twelve social categories reveals that overall valence fluctuations undergo a distinct layer-wise smoothing, characterized by a significant hierarchical representation divergence as the layers deepen, which ultimately manifests as a widespread mitigation of bias at the behavioral level. In stark contrast to this macro-level reduction, this reflection mechanism is not universally corrective, instead exhibiting a stubborn, category-specific selectivity that regularly locks in and perversely amplifies localized biases. Warning: this paper contains examples with biased content.

cs.SI↗

arXiv2026

Modeling Implicit Conflict Monitoring Mechanisms against Stereotypes in LLMs

In this paper, we study an emergent self-debiasing mechanisms against stereotypical content in Large Language Models (LLMs). Unlike traditional safety mechanisms that are primarily triggered by explicit input-level stimuli, self-debiasing mechanisms can involve generation-time intrinsic correction that are not directly reducible to surface-level prompt. Motivated by conflict-monitoring and response-inhibition accounts in cognitive neuroscience, we propose COCO, a contrastive causal method designed to identify COCO neurons that exhibit high intra-\underline{CO}nsistency yet sharp inter-\underline{CO}ntrast across antithetical generative responses, such as stereotypical versus unbiased outputs. Ablation studies reveal that deactivating COCO neurons leads to a catastrophic collapse of the model's fairness; over 90\% of outputs revert to biased content, far exceeding the bias levels induced by explicit adversarial jailbreak attacks. Observing that simple weight amplification of COCO neurons yields only marginal gains, we propose two training-free, lightweight editing strategies: Local Enhancement (LE-COCO) and Networked Enhancement (NE-COCO). Comprehensive evaluations show that our methods bolster robustness against adversarial jailbreaks and achieve strong performance on open-ended safety benchmarks, while preserving foundational generative proficiency. While this study primarily addresses social stereotypes, the COCO mechanism holds significant potential for diverse domains like hallucination detection, offering valuable insights toward the development of self-evolving AI agents.

cs.SI↗

arXiv2025

Approximation via partial Hausdorff integrals on $H^1(\mathbb{R})$

We obtain the result of approximating $ f $ in the $ H^1(\mathbb{R}) $ norm using partial Hausdorff integrals. Specifically, by leveraging the homogeneous multiplier theory of $ H^1(\mathbb{R}) $ and the $ K $ functional theory, one result from Pinos and Liflyand [CMB,~2021,~64,~no.3] is extended from $ L^p(\mathbb{R}) $ ( $ 1 \leq p \leq \infty $) to $ H^1(\mathbb{R}) $. As applications, four examples of partial Hausdorff integrals are also given.

math.CA↗

arXiv2025

Estimates of fractional Hausdorff operators on weighted Lebesgue and Hardy spaces

In this article, we obtain some necessary and sufficient conditions for the boundedness of fractional Hausdorff operators $h_{\Phi,\beta}$ on weighted Lebesgue spaces $(0\leq\beta<1)$, which are fractional variants of Bandaliev-Safarova [Hacet. J. Math. Stat., 2021, 50, 1334-1346]; it is found that a new constraint for $\beta$ should be added and it holds automatically for non-fractional variants in [HJMS, 2021] $(\beta=0)$. Then, we further obatin the boundedness of fractional Hausdorff operators $h_{\Phi,\beta}$ on power-weighted Hardy spaces which are fractional variants of Ruan-Fan [Math. Nachr., 2017, 290, 2388-2400]. Ruan-Fan obtained the relevant boundedness results by means of the radial maximal function characterization of Hardy spaces, while in this paper, two different relevant results are obtained respectively by using the radial maximal function characterization and the Riesz characterization of power-weighted Hardy spaces.

math.CA↗

arXiv2025

A Multi-Memory Segment System for Generating High-Quality Long-Term Memory Content in Agents

In the current field of agent memory, extensive explorations have been conducted in the area of memory retrieval, yet few studies have focused on exploring the memory content. Most research simply stores summarized versions of historical dialogues, as exemplified by methods like A-MEM and MemoryBank. However, when humans form long-term memories, the process involves multi-dimensional and multi-component generation, rather than merely creating simple summaries. The low-quality memory content generated by existing methods can adversely affect recall performance and response quality. In order to better construct high-quality long-term memory content, we have designed a multi-memory segment system (MMS) inspired by cognitive psychology theory. The system processes short-term memory into multiple long-term memory segments, and constructs retrieval memory units and contextual memory units based on these segments, with a one-to-one correspondence between the two. During the retrieval phase, MMS will match the most relevant retrieval memory units based on the user's query. Then, the corresponding contextual memory units is obtained as the context for the response stage to enhance knowledge, thereby effectively utilizing historical data. We conducted experiments on the LoCoMo dataset and further performed ablation experiments, experiments on the robustness regarding the number of input memories, and overhead experiments, which demonstrated the effectiveness and practical value of our method.

cs.AI↗