arXiv AI
Aug 5
Instruction Stacking Collapse: A Benchmark and the Capability-Dependent Value of Prompt Compilation
arXiv:2608. 02639v1 Announce Type: cross Abstract: Production prompts rarely carry a single instruction.
Knowledge-Conditioned, Single-Pass LLM Synthesis of Executable Unity Game Scenes: A Compiler Error Census across 26 Goal Playable Concepts
arXiv:2607. 10187v1 Announce Type: new Abstract: Large language models (LLMs) write Unity C\# for game scenes.
Related stories
arXiv AI
Jun 2
Accuracy, Stability, and Repeated-Run Reliability of Large Language Models on Deterministic Programming Tasks
arXiv:2606. 00920v1 Announce Type: cross Abstract: Run-level pass rate overstates retry-free coverage by up to 17.
arXiv AI
Aug 18
ReLoop: Structured Modeling and Behavioral Verification for Reliable LLM-Based Optimization
arXiv:2602. 15983v3 Announce Type: replace-cross Abstract: Large language models (LLMs) can translate natural language into optimization code, but silent failures pose a critical risk: code that executes and returns solver-feasible solutions may encode semantically incorrect formulations---a feasibility--correctness gap reaching 90 percentage points on compositional problems.
arXiv AI
Aug 14
Large Language Models Can Follow Instructions, But Not Many at Once: Phase Transitions in Compositional Constraint Satisfaction
arXiv:2608. 12426v1 Announce Type: new Abstract: Large language models are increasingly deployed in settings that require simultaneous adherence to multiple explicit constraints - reasoning structure, safety boundaries, output schemas.
arXiv AI
Jun 2
How Generation Architecture Shapes Code Complexity in Multi-Agent LLM Systems: A Paired Study on HumanEval
arXiv:2606. 00308v1 Announce Type: cross Abstract: Large-language-model code generation has shifted from single-shot prompting to multi-agent orchestrations - analyst, coder, tester, and debugger pipelines - and is evaluated almost exclusively on functional correctness.
arXiv Machine Learning
3d ago
Narcissus: Program Synthesis Using Context-Aware LLM Approximations
arXiv:2608.25657v1 Announce Type: cross Abstract: Large language models (LLMs) excel at programming, but not when the task fixes the target language: prompted with a grammar rare in their training da...
arXiv AI
Jun 9
Where Instruction Hierarchy Breaks: Diagnosing and Repairing Failures in Reasoning Language Models
arXiv:2606. 07808v1 Announce Type: new Abstract: Reasoning language models deployed in agentic workflows must follow an instruction hierarchy: when instructions from different sources conflict, the model should obey the highest-privilege applicable instruction.
arXiv AI
5d ago
Architecture as Capability Equalizer for Coding Agents
arXiv:2608.21747v1 Announce Type: cross Abstract: LLM-based coding agents generate complete software systems from high-level descriptions, yet little is known about how the format of architecture spe...
arXiv AI
Jul 7
Obey, Diverge, Collapse: Blind Obedience to Incorrect Instructions Drives Code LLMs to Irrecoverable Code Semantic Collapse
arXiv:2607. 04537v1 Announce Type: cross Abstract: Code language models are now trusted collaborators in production workflows for debugging, refactoring, and iterative repair, and every benchmark that evaluates them assumes the instructions they act on are correct.
arXiv AI
Aug 5
TraceCompiler: Skill-Guided Mining and Compilation of LLM Agent Traces into Mostly Deterministic Workflows
arXiv:2608. 02680v1 Announce Type: cross Abstract: Tool-using language-model agents repeatedly rediscover procedures they have already executed, producing traces that mix reusable structure with retries, exploration, accidental ordering, and repeated lookups.
arXiv AI
Jun 9
Lost in the Flow with Code Talkers: Unveiling the Instruction-Tuning Tax of Large Language Models in Code Tasks
arXiv:2606. 08676v1 Announce Type: cross Abstract: AI coding assistants have significantly improved developer productivity by automatically suggesting code that aligns with user intent, and many of these tools are now integrated directly into Integrated Development Environments (IDEs).
arXiv AI
Aug 11
Revision or Re-Solving? Decomposing Second-Pass Gains in Multi-LLM Pipelines
arXiv:2604. 01029v2 Announce Type: replace-cross Abstract: Multi-LLM revision pipelines, in which a second model reviews and improves a draft produced by a first, are widely assumed to derive their gains from genuine error correction.