Temperature, Not Knowledge: What Recursive Self-Training and Its Remedies Do to a Language Model's Confidence
cs.AI
Submitted: 2026-10-31
Updated: 2026-10-31
Terminology
Sources
- Top-H Decoding: Adapting the Creativity and Coherence with Bounded Entropy in Text Generation
- Beyond Model Collapse: Scaling Up with Synthesized Data Requires Verification
- Is Model Collapse Inevitable? Breaking the Curse of Recursion by Accumulating Real and Synthetic Data
- Beyond Accuracy: The Role of Calibration in Self-Improving Large Language Models
- Language Models (Mostly) Know What They Know
- Your Pre-trained LLM is Secretly an Unsupervised Confidence Calibrator
- When Sample Selection Bias Precipitates Model Collapse
- Qwen3 Technical Report
- Continual Calibration: Coverage Can Collapse Before Accuracy in Lifelong LLM Fine-Tuning
- Escaping Model Collapse via Synthetic Data Verification: Near-term Improvements and Long-term Convergence
- ForTIFAI: Fending Off Recursive Training Induced Failure for AI Model Collapse
Related papers
- Multi-Objective Structured Pruning of LLMs for Latency and Model Size Optimization
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Embeddings for Preferences, Not Semantics
- Can We Trust Item Response Theory for AI Evaluation?
- Diagnosing Task Insensitivity in Language Agents