Papers with three-stage pipeline
Compilable Neural Code Generation with Compiler Feedback (2022.findings-acl)
Copied to clipboard
Xin Wang, Yasheng Wang, Yao Wan, Fei Mi, Yitong Li, Pingyi Zhou, Jin Liu, Hao Wu, Xin Jiang, Qun Liu
| Challenge: | Existing deep-learning approaches model code generation as text generation, but few of them account for compilability of the generated programs. |
| Approach: | They propose a three-stage pipeline utilizing compiler feedback for compilable code generation to improve compilability. |
| Outcome: | The proposed pipeline improves compilability of generated programs by combining compiler feedback, language model fine-tuning, and compilable discrimination. |
A New Pipeline for Knowledge Graph Reasoning Enhanced by Large Language Models Without Fine-Tuning (2024.emnlp-main)
Copied to clipboard
| Challenge: | Conventional knowledge Graph Reasoning models learn the embeddings of KG components over the structure of a KG. |
| Approach: | They propose a pipeline to integrate knowledge from LLMs into KGs without fine-tuning . they propose knowledge alignment, KG reasoning and entity reranking to enhance conventional models . |
| Outcome: | The proposed pipeline can enhance the performance of conventional KGR models in incomplete and general situations. |
RAVEN: Query-Guided Representation Alignment for Question Answering over Audio, Video, Embedded Sensors, and Natural Language (2025.emnlp-main)
Copied to clipboard
| Challenge: | Multimodal question answering often requires identifying which video, audio, or sensor tokens are relevant to the question. off-camera speech, background noise, or motion outside the field of view often mislead fusion models that weight all streams equally. |
| Approach: | They propose a unified architecture for multimodal question answering that assigns scalar relevance scores to each token across modalities. |
| Outcome: | The proposed model outperforms state-of-the-art multimodal large language models on seven multi-modal QA benchmarks and egocentric and exocentric tasks. |
Towards Self-Improving Error Diagnosis in Multi-Agent Systems (2026.findings-acl)
Copied to clipboard
| Challenge: | Existing diagnostic approaches rely on expensive expert annotations and ”LLM-as-a-judge” paradigms. |
| Approach: | They propose a framework for semantic failure attribution that identifies responsible agents and the originating error step. |
| Outcome: | The proposed framework outperforms baselines in step-level localization and validation. |
General-to-Specific Transfer Labeling for Domain Adaptable Keyphrase Generation (2023.findings-acl)
Copied to clipboard
| Challenge: | Large distribution shifts among different domains hinder transferability of keyphrase generation models. |
| Approach: | They propose a pipeline which guides KPG models’ learning focus from general syntactical features to domain-related semantics in a data-efficient manner. |
| Outcome: | The proposed pipeline can produce good quality keyphrases in new domains and achieve consistent improvements after adaptation with limited in-domain annotated data. |
SKA-Bench: A Fine-Grained Benchmark for Evaluating Structured Knowledge Understanding of LLMs (2025.findings-emnlp)
Copied to clipboard
| Challenge: | Existing evaluations for Structured Knowledge (SK) understanding are non-rigorous and focus on a single type of SK. |
| Approach: | They propose a structured knowledge understanding benchmark that includes four widely used structured knowledge forms. |
| Outcome: | The proposed benchmark is based on four widely used structured knowledge forms . it includes a question, an answer, positive knowledge units, and noisy knowledge units . |
Graph-Based Chain-of-Thought Pruning for Reducing Redundant Reflections in Reasoning LLMs (2026.findings-acl)
Copied to clipboard
Hongyuan Yuan, Xinran He, Run Shao, Bolei He, Xianwei Xue, Mengke Chen, Qiutong Pan, Haiwei Wang, Haifeng Li
| Challenge: | Extending CoT through RL can induce undesirable thinking patterns such as overthinking . prior work has focused on inefficient reflection, which manifests in two problematic patterns: Indiscriminate Reflection and Repetitive Reflectione . |
| Approach: | They propose a graph-based approach to optimize CoT by pruning each linear CoT into a directed acyclic graph with explicit dependency edges. |
| Outcome: | The proposed approach reduces the average reasoning tokens by 42% while maintaining or improving accuracy. |
CoReTab: Improving Multimodal Table Understanding with Code-driven Reasoning (2026.eacl-long)
Copied to clipboard
| Challenge: | Existing datasets for multimodal table understanding provide short factual answers without explicit multi-step reasoning supervision. |
| Approach: | They propose a code-driven reasoning framework that produces scalable, interpretable, and automatically verifiable annotations by coupling multi-step reasoning with executable Python code. |
| Outcome: | The proposed model achieves significant gains over baseline models while producing transparent and verifiable reasoning traces. |
NILE: Internal Consistency Alignment in Large Language Models (2025.emnlp-main)
Copied to clipboard
Minda Hu, Qiyuan Zhang, Yufei Wang, Bowei He, Hongru Wang, Jingyan Zhou, Liangyou Li, Yasheng Wang, Chen Ma, Irwin King
| Challenge: | Recent advances show that the world knowledge in the Instruction Fine-Tuning (IFT) dataset, which is incompatible with LLMs’ internal knowledge, can greatly hurt the IFT performance. |
| Approach: | They propose a framework to optimize the effectiveness of IFT by carefully aligning the world and internal knowledge of LLMs. |
| Outcome: | The proposed framework can significantly improve performance across multiple LLM ability evaluation datasets. |
SafeMCP: Proactive Power Regulation for LLM Agent Defense via Environment-Grounded Look-Ahead Reasoning (2026.acl-long)
Copied to clipboard
| Challenge: | Large Language Model (LLM) agents are expanding their action spaces to operate in complex environments. |
| Approach: | They propose a server-side defense plugin that constrains tool acquisition via predictive reasoning regarding future safety risks. |
| Outcome: | Experiments on PowerSeeking Bench, ToolEmu, and AgentHarm show that SafeMCP achieves a safe equilibrium, effectively mitigating risks while preserving agent utility. |
Balancing Fidelity and Plasticity: Aligning Mixed-Precision Fine-Tuning with Linguistic Hierarchies (2026.findings-acl)
Copied to clipboard
Changhai Zhou, Shiyang Zhang, Yuhua Zhou, Jun Gao, Qian Qiao, Shichao Weng, Weizhong Zhang, Cheng Jin
| Challenge: | Existing quantization-aware fine-tuning methods decouple weight precision and adapter capacity, overlooking that a layer’s ability to adapt is constrained by the information preserved in its frozen weights. |
| Approach: | They propose a framework that jointly optimizes per-layer quantization bit-width and LoRA rank. |
| Outcome: | Experiments on LLaMA and Qwen models show that the proposed framework matches or approaches 16-bit baselines while using substantially less memory. |
NASH: Numerically Aware Scoring Heuristic for Robust Semantic Similarity (2026.findings-acl)
Copied to clipboard
| Challenge: | Numerical precision is critical in financial NLP, yet embedding-based semantic similarity metrics exhibit numerical blindness. |
| Approach: | They propose a model-agnostic metric that decouples numerical verification from textual semantic evaluation. |
| Outcome: | The proposed metric improves numerical sensitivity while maintaining general semantic performance. |
RoleConflictBench: A Benchmark of Role Conflict Scenarios for Evaluating LLMs’ Contextual Sensitivity (2026.findings-acl)
Copied to clipboard
| Challenge: | a new benchmark measures the contextual sensitivity of large language models in role conflict scenarios . role conflicts are social dilemmas where multiple roles cannot be fulfilled simultaneously . authors: models are forced to arbitrate between dynamic contextual cues and learned preferences . |
| Approach: | They propose a benchmark to measure the contextual sensitivity of large language models in role conflict scenarios. |
| Outcome: | The proposed benchmark measures the contextual sensitivity of large language models in role conflict scenarios. |
MirrorQA: Benchmarking Multimodal LLMs on Mirror-Orientation Reasoning (2026.acl-long)
Copied to clipboard
| Challenge: | Multimodal large language models (MLLMs) have achieved remarkable progress in recent years, yet their ability to perform left–right reasoning in mirror contexts remains underexplored. |
| Approach: | They propose a benchmark to evaluate MLLMs' ability to distinguish left from right from a subject-centered perspective. |
| Outcome: | The proposed benchmarks show that even the best performing models achieve only 65.40% accuracy, far below the 99.28% accuracy of humans. |
GenPT: Beyond Self-Report for Reliable LLM Psychometrics via Generative Projective Testing (2026.acl-long)
Copied to clipboard
Ming Wang, Shuang Wu, Bixuan Wang, Lu Lin, Yuxin Chen, Xiaocui Yang, Daling Wang, Shi Feng, Yifei Zhang, Yufan Sun
| Challenge: | Large language models (LLMs) inherit contamination from training corpora, directional bias under social-desirability framing, and limited responsiveness to context beyond the item text. |
| Approach: | They propose a paradigm that reformulates TAT, Rorschach, and SCT with newly generated stimuli and organises assessment as a three-stage pipeline. |
| Outcome: | The proposed paradigm reformulates TAT, Rorschach, and SCT with newly generated stimuli and organises assessment as a three-stage pipeline. |