Can Large Language Models Still Explain Themselves? Investigating the Impact of Quantization on Self-Explanations

Name: Can Large Language Models Still Explain Themselves? Investigating the Impact of Quantization on Self-Explanations
Uploaded: 2026-08-26T02:08:21.056615+00:00
Duration: 482 s
Description: The episode discusses a paper investigating how model quantization—making large language models smaller—affects their ability to self-explain. Hosts conclude that while quantization improves efficiency, it systematically degrades the fidelity and quality of explanations, suggesting explainability must be architecturally protected.

The episode discusses a paper investigating how model quantization—making large language models smaller—affects their ability to self-explain. Hosts conclude that while quantization improves efficiency, it systematically degrades the fidelity and quality of explanations, suggesting explainability must be architecturally protected.

Open the full episode