Tootfinder

Assessing and Mitigating Data Memorization Risks in Fine-Tuned Large Language Models
Badrinath Ramakrishnan, Akshaya Balaji
https://arxiv.org/abs/2508.14062 https://

Assessing and Mitigating Data Memorization Risks in Fine-Tuned Large Language Models
Large Language Models (LLMs) have demonstrated remarkable capabilities across diverse natural language processing tasks, but their tendency to memorize training data poses significant privacy risks, particularly during fine-tuning processes. This paper presents a comprehensive empirical analysis of data memorization in fine-tuned LLMs and introduces a novel multi-layered privacy protection framework. Through controlled experiments on modern LLM architectures including GPT-2, Phi-3, and Gemma-2,…

MTMD: A Multi-Task Multi-Domain Framework for Unified Ad Lightweight Ranking at Pinterest
Xiao Yang, Peifeng Yin, Abe Engle, Jinfeng Zhuang, Ling Leng
https://arxiv.org/abs/2510.09857

MTMD: A Multi-Task Multi-Domain Framework for Unified Ad Lightweight Ranking at Pinterest
The lightweight ad ranking layer, living after the retrieval stage and before the fine ranker, plays a critical role in the success of a cascaded ad recommendation system. Due to the fact that there are multiple optimization tasks depending on the ad domain, e.g., Click Through Rate (CTR) for click ads and Conversion Rate (CVR) for conversion ads, as well as multiple surfaces where an ad is served (home feed, search, or related item recommendation) with diverse ad products (shopping or standard…

TOUCAN: Synthesizing 1.5M Tool-Agentic Data from Real-World MCP Environments
Zhangchen Xu, Adriana Meza Soria, Shawn Tan, Anurag Roy, Ashish Sunil Agrawal, Radha Poovendran, Rameswar Panda
https://arxiv.org/abs/2510.01179

TOUCAN: Synthesizing 1.5M Tool-Agentic Data from Real-World MCP Environments
Large Language Model (LLM) agents are rapidly emerging as powerful systems for automating tasks across domains. Yet progress in the open-source community is constrained by the lack of high quality permissively licensed tool-agentic training data. Existing datasets are often limited in diversity, realism, and complexity, particularly regarding multi-tool and multi-turn interactions. To address this gap, we introduce Toucan, the largest publicly available tool-agentic dataset to date, containing …

An underdog story: Re-emergence of a polar instability at high pressure in KNbO3
Mohamad Baker Shoker, Sitaram Ramakrishnan, Boris Croes, Olivier Cregut, Nicolas Beyer, Kokou Dorkenoo, Pierre Rodi\`ere, Bj\"orn Wehinger, Gaston Garbarino, Mohamed Mezouar, Marine Verseils, Pierre Fertey, Salia Cherifi-Hertel, Pierre Bouvier, Mael Guennou
https://

An underdog story: Re-emergence of a polar instability at high pressure in KNbO3
Ferroelectricity in perovskites is known to be suppressed by a moderate hydrostatic pressure. The notion that a polar instability should reappear in a higher pressure regime is well accepted theoretically but experiments have failed so far to provide a conclusive evidence for it. Here, we investigate a classical but comparatively underlooked ferroelectric perovskite KNbO3. We use single crystal X-ray diffraction, infrared and Raman spectroscopy and second-harmonic generation to explore the phas…

Emergent Rashba spin-orbit coupling in bulk gold with buried network of nanoscale interfaces
Shreya Kumbhakar, Banashree Debnath, Tuhin Kumar Maji, Binita Tongbram, Shinjan Mandal, T. Phanindra Sai, T. V. Ramakrishnan, Manish Jain, H. R. Krishnamurthy, Anshu Pandey, Arindam Ghosh
https://arxiv.org/abs/2509.03620

Emergent Rashba spin-orbit coupling in bulk gold with buried network of nanoscale interfaces
The Rashba effect, which plays a crucial role in fundamental materials physics and potential spintronics applications, has been engineered in diverse systems, including semiconductor quantum wells, oxide heterostructures, metallic surfaces, topological insulators, ferroelectrics, etc. However, generating it in systems that preserve bulk inversion symmetry (BIS), for example, in bulk metals, has not been possible so far. We demonstrate a unique strategy to introduce and tune Rashba spin-orbit in…

Language Bias in Information Retrieval: The Nature of the Beast and Mitigation Methods
Jinrui Yang, Fan Jiang, Timothy Baldwin
https://arxiv.org/abs/2509.06195 https://

Language Bias in Information Retrieval: The Nature of the Beast and Mitigation Methods
Language fairness in multilingual information retrieval (MLIR) systems is crucial for ensuring equitable access to information across diverse languages. This paper sheds light on the issue, based on the assumption that queries in different languages, but with identical semantics, should yield equivalent ranking lists when retrieving on the same multilingual documents. We evaluate the degree of fairness using both traditional retrieval methods, and a DPR neural ranker based on mBERT and XLM-R. A…