@misc{indiciaee8d3161d8967,
  title = {From Unreliable to Usable: Decomposing Failure Modes and Confidence Signals in LLM Entity Extraction - Evidence from Cyber Threat Intelligence},
  author = {Jiahong Qian},
  year = {2026},
  url = {https://arxiv.org/abs/2610.20417},
  note = {Source identifier: 2610.20417}
}