Copertina

Asymmetric Goal Drift in Coding Agents Under Value Conflict

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Saebo, Magnus, Gibson, Spencer, Crosse, Tyler, Menon, Achyutha, Jang, Eyon, Cruz, Diogo
Natura:	Preprint
Pubblicazione:	2026
Soggetti:	Artificial Intelligence Computation and Language Software Engineering
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

Inherited Goal Drift: Contextual Pressure Can Undermine Agentic Goals
di: Menon, Achyutha, et al.
Pubblicazione: (2026)

Value-Conflict Diagnostics Reveal Widespread Alignment Faking in Language Models
di: Nair, Inderjeet, et al.
Pubblicazione: (2026)

Exploring Language Model's Code Generation Ability with Auxiliary Functions
di: Lee, Seonghyeon, et al.
Pubblicazione: (2024)

SWE-Spot: Building Small Repo-Experts with Repository-Centric Learning
di: Peng, Jinjun, et al.
Pubblicazione: (2026)

Chain of Grounded Objectives: Bridging Process and Goal-oriented Prompting for Code Generation
di: Yeo, Sangyeop, et al.
Pubblicazione: (2025)

Eliciting Instruction-tuned Code Language Models' Capabilities to Utilize Auxiliary Function for Code Generation
di: Lee, Seonghyeon, et al.
Pubblicazione: (2024)

LocAgent: Graph-Guided LLM Agents for Code Localization
di: Chen, Zhaoling, et al.
Pubblicazione: (2025)

CodeScout: An Effective Recipe for Reinforcement Learning of Code Search Agents
di: Sutawika, Lintang, et al.
Pubblicazione: (2026)

Polygon: Symbolic Reasoning for SQL using Conflict-Driven Under-Approximation Search
di: Zhao, Pinhan, et al.
Pubblicazione: (2025)

Drift-Bench: Diagnosing Cooperative Breakdowns in LLM Agents under Input Faults via Multi-Turn Interaction
di: Bao, Han, et al.
Pubblicazione: (2026)

Training Versatile Coding Agents in Synthetic Environments
di: Zhu, Yiqi, et al.
Pubblicazione: (2025)

Effective Harness Engineering for Algorithm Discovery with Coding Agents
di: Ishibashi, Yoichi, et al.
Pubblicazione: (2026)

AInsteinBench: Benchmarking Coding Agents on Scientific Repositories
di: Duston, Titouan, et al.
Pubblicazione: (2025)

Dynamic Benchmarking of Reasoning Capabilities in Code Large Language Models Under Data Contamination
di: Chen, Simin, et al.
Pubblicazione: (2025)

How Diversely Can Language Models Solve Problems? Exploring the Algorithmic Diversity of Model-Generated Code
di: Lee, Seonghyeon, et al.
Pubblicazione: (2025)

Human-Agent Collaborative Paper-to-Page Crafting for Under $0.1
di: Ma, Qianli, et al.
Pubblicazione: (2025)

SlopCodeBench: Benchmarking How Coding Agents Degrade Over Long-Horizon Iterative Tasks
di: Orlanski, Gabriel, et al.
Pubblicazione: (2026)

Can Coding Agents Reproduce Findings in Computational Materials Science?
di: Huang, Ziyang, et al.
Pubblicazione: (2026)

OmniCode: A Benchmark for Evaluating Software Engineering Agents
di: Sonwane, Atharv, et al.
Pubblicazione: (2026)

CodeR: Issue Resolving with Multi-Agent and Task Graphs
di: Chen, Dong, et al.
Pubblicazione: (2024)

Code Broker: A Multi-Agent System for Automated Code Quality Assessment
di: Attrah, Samer
Pubblicazione: (2026)

SpecBench: Measuring Reward Hacking in Long-Horizon Coding Agents
di: Zhao, Bingchen, et al.
Pubblicazione: (2026)

MOSS: Enabling Code-Driven Evolution and Context Management for AI Agents
di: Zhu, Ming, et al.
Pubblicazione: (2024)

CRANE: Constrained Reasoning Injection for Code Agents via Nullspace Editing
di: Zhu, Mingzhi, et al.
Pubblicazione: (2026)

TokDrift: When LLM Speaks in Subwords but Code Speaks in Grammar
di: Li, Yinxi, et al.
Pubblicazione: (2025)

SWE-Chain: Benchmarking Coding Agents on Chained Release-Level Package Upgrades
di: Lam, Man Ho, et al.
Pubblicazione: (2026)

CR-Bench: Evaluating the Real-World Utility of AI Code Review Agents
di: Pereira, Kristen, et al.
Pubblicazione: (2026)

CODESIM: Multi-Agent Code Generation and Problem Solving through Simulation-Driven Planning and Debugging
di: Islam, Md. Ashraful, et al.
Pubblicazione: (2025)

VisCoder2: Building Multi-Language Visualization Coding Agents
di: Ni, Yuansheng, et al.
Pubblicazione: (2025)

Co-Learning: Code Learning for Multi-Agent Reinforcement Collaborative Framework with Conversational Natural Language Interfaces
di: Yu, Jiapeng, et al.
Pubblicazione: (2024)

Evaluating LLM-Based Goal Extraction in Requirements Engineering: Prompting Strategies and Their Limitations
di: Arnaudo, Anna, et al.
Pubblicazione: (2026)

Library Drift: Diagnosing and Fixing a Silent Failure Mode in Self-Evolving LLM Skill Libraries
di: Zhang, Xing, et al.
Pubblicazione: (2026)

EvoCodeBench: An Evolving Code Generation Benchmark Aligned with Real-World Code Repositories
di: Li, Jia, et al.
Pubblicazione: (2024)

IndustryCode: A Benchmark for Industry Code Generation
di: Zeng, Puyu, et al.
Pubblicazione: (2026)

Rethinking Code Refinement: Learning to Judge Code Efficiency
di: Seo, Minju, et al.
Pubblicazione: (2024)

CodeClash: Benchmarking Goal-Oriented Software Engineering
di: Yang, John, et al.
Pubblicazione: (2025)

OpenCodeInterpreter: Integrating Code Generation with Execution and Refinement
di: Zheng, Tianyu, et al.
Pubblicazione: (2024)

CodeMirage: Hallucinations in Code Generated by Large Language Models
di: Agarwal, Vibhor, et al.
Pubblicazione: (2024)

A Code Comprehension Benchmark for Large Language Models for Code
di: Havare, Jayant, et al.
Pubblicazione: (2025)

CodeJudgeBench: Benchmarking LLM-as-a-Judge for Coding Tasks
di: Jiang, Hongchao, et al.
Pubblicazione: (2025)