Cross-lingual Calibration of Pre-Generation Success Probes for Multilingual LLM Routing
cs.CL, cs.AI
Submitted: 2026-10-05
Updated: 2026-10-05
Terminology
Sources
- LLMs Encode Relevance as a Layer-Wise Cross-Lingual Signal
- No Answer Needed: Predicting LLM Answer Accuracy from Question-Only Linear Probes
- FrugalGPT: How to Use Large Language Models While Reducing Cost and Improving Performance
- Do LLM hallucination detectors suffer from low-resource effect?
- Code Correctness Is Linearly Decodable from LLM Hidden States Before Generation
- Universal Model Routing for Efficient LLM Inference
- How Language Directions Align with Token Geometry in Multilingual LLMs
- LLMs Encode Their Failures: Predicting Success from Pre-Generation Activations
- LLMs Encode How Difficult Problems Are
- RouteLLM: Learning to Route LLMs with Preference Data
- gpt-oss-120b & gpt-oss-20b Model Card
- Language Models are Multilingual Chain-of-Thought Reasoners
- LLM Router: Rethinking Routing with Prefill Activations
- Qwen3 Technical Report
Related papers
- Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
- Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
- Subliminal Steering: Stronger Encoding of Hidden Signals
- MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
- The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
- Untangling the Mechanisms of Misleading Context in Medical Question Answering