Yu Liang, Liangxin Liu, Longzheng Wang, Wangyan, Zhang Yueyang, Long Xia, Zhiyuan Sun, Daiting Shi. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Kai Qin, Liangxin Liu, Yu Liang, Longzheng Wang, Wangyan, Zhang Yueyang, Long Xia, Zhiyuan Sun, Houde Liu, Daiting Shi. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

The Ethereum ecosystem, which secures over $381 billion in assets, fundamentally relies on client APIs as the sole interface between users and the blockchain. However, these critical APIs suffer from widespread implementation inconsistencies, which can lead to financial discrepancies, degraded user experiences, and threats to network reliability. Despite this criticality, existing testing approaches remain manual and incomplete: they require extensive domain expertise, struggle to keep pace with Ethereum’s rapid evolution, and fail to distinguish genuine bugs from acceptable implementation variations. We present APIDiffer , the first specification-guided differential testing framework designed to automatically detect API inconsistencies across Ethereum’s diverse client ecosystem. APIDiffer transforms API specifications into comprehensive test suites through two key innovations: (1) specification-guided test input generation that creates both syntactically valid and invalid requests enriched with real-time blockchain data, and (2) specification-aware false positive filtering that leverages large language models to distinguish genuine bugs from acceptable variations. Our evaluation across all 11 major Ethereum clients reveals the pervasiveness of API bugs in production systems. APIDiffer uncovered 72 bugs, with 90.28% already confirmed or fixed by developers, including one critical error in the official specifications themselves. Beyond these raw numbers, APIDiffer achieves up to 89.67% higher code coverage than existing tools and reduces false positive rates by 37.38%. The Ethereum community’s response validates our impact: developers have integrated our test cases, expressed interest in adopting our methodology, and escalated one bug to the official Ethereum Project Management meeting. By making APIDiffer open-source, we enable continuous validation of Ethereum client API implementations, thereby strengthening the foundational integrity of the entire Ethereum ecosystem.

Consistency learning (CL) has proven to be a valuable technique for improving the robustness of models in conditional sentence generation (CSG) tasks by ensuring stable predictions across various input data forms.However, models augmented with CL often face challenges in optimizing consistency features, which can detract from their efficiency and effectiveness.To address these challenges, we introduce Curriculum Consistency Learning (CCL), a novel strategy that guides models to learn consistency in alignment with their current capacity to differentiate between features.CCL is designed around the inherent aspects of CL-related losses, promoting task independence and simplifying implementation.Implemented across four representative CSG tasks, including instruction tuning (IT) for large language models and machine translation (MT) in three modalities (text, speech, and vision), CCL demonstrates marked improvements.Specifically, it delivers +2.0 average accuracy point improvement compared with vanilla IT and an average increase of +0.7 in COMET scores over traditional CL methods in MT tasks.Our comprehensive analysis further indicates that models utilizing CCL are particularly adept at managing complex instances, showcasing the effectiveness and efficiency of CCL in improving CSG models.Code and scripts are available at https://github.com/xinxinxing/ Curriculum-Consistency-Learning.

Instruction tuning (IT) is crucial to tailoring large language models (LLMs) towards human-centric interactions. Recent advancements have shown that the careful selection of a small, high-quality subset of IT data can significantly enhance the performance of LLMs. Despite this, common approaches often rely on additional models or data, which increases costs and limits widespread adoption. In this work, we propose a novel approach, termed $\textit{SelectIT}$, that capitalizes on the foundational capabilities of the LLM itself. Specifically, we exploit the intrinsic uncertainty present in LLMs to more effectively select high-quality IT data, without the need for extra resources. Furthermore, we introduce a curated IT dataset, the $\textit{Selective Alpaca}$, created by applying SelectIT to the Alpaca-GPT4 dataset. Empirical results demonstrate that IT using Selective Alpaca leads to substantial model ability enhancement. The robustness of SelectIT has also been corroborated in various foundation models and domain-specific tasks. Our findings suggest that longer and more computationally intensive IT data may serve as superior sources of IT, offering valuable insights for future research in this area. Data, code, and scripts are freely available at https://github.com/Blue-Raincoat/SelectIT.

Summary Raft is a consensus algorithm that implements highly available replicas. As raft requires a leader to drive the protocol execution, it usually employs a round‐robin leader election mechanism to select a random leader among all nodes. However, when the elected leader suffers from performance issues, the protocol will temporarily lose aliveness. To address this issue, we propose a multi‐strategy leader election mechanism that allows nodes to proactively trigger a new round of election when they believe the leader suffers from performance issues, and replace this leader with a new one in a designated priority queue. The formal analysis and experimental results show that the mechanism can improve the system's throughput/latency without additional election time overhead.

State machine replication (SMR) allows nodes to jointly maintain a consistent ledger, even when a part of nodes are Byzantine. To defend against and/or limit the impact of attacks launched by Byzantine nodes, there have been blocks that combine reputation mechanisms to SMR, where each node has a reputation value based on its historical behaviours, and the node’s voting power will be proportional to its reputation. Despite the promising features of reputation-based SMR, existing studies do not provide formal treatment on the reputation mechanism on SMR protocols, including the types of behaviours affecting the reputation, the security properties of the reputation mechanism, and the extra security properties of SMR using reputation mechanisms.In this paper, we provide the first formal study on the reputation-based SMR. We define the security properties of the reputation mechanism w.r.t. these misbehaviours. Based on the formalisation of the reputation mechanism, we formally define the reputation-based SMR, and identify a new property reputation-consistency that is necessary for ensuring reputation-based SMR’s safety. We then design a simple reputation mechanism that achieves all security properties in our formal model. To demonstrate the practicality, we combine our reputation mechanism to the Sync-HotStuff SMR protocol, yielding a simple and efficient reputation-based SMR at the cost of only an extra ∆ in latency, where ∆ is the maximum delay in synchronous networks.