Unverified50% confidenceBenchmarkExact time
On the frozen test set, the delivered system (LoRA + rules) caught 73 (recall 0.86), with 11 confident errors, 1 "not sure", and 0 false positives, outperforming the keyword baseline, LoRA-only fine-tuning, and zero-shot Qwen3.5-4B.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/18/2026
First Seen
Valid until: 1/16/2027
Sources
Related Entities
Related Claims
UnverifiedLargha-150M-v1's training mixture includes a pool of 500 error corrections, which were mined from the model's predictions on typed-decision test rows, so 0.906 is a fitting reading rather than a zero-shot result.75% similarUnverifiedOn a test set of 600 synthetic invoices, the QLoRA fine-tuned model achieved a full-invoice exact rate of 83.0%, an auto-approval rate of 86.8%, and 0 out of 461 validation-coverage field errors in approved invoices (upper bound 0.8%).72% similarUnverifiedRanked by question difficulty, persistence rate was 91.4% on the 60 questions that all models answered correctly, but only 73.5% on the 52 questions that at least one model got wrong, indicating that sycophancy tracks uncertainty.72% similarUnverifiedWhen the open-source small model gemma2:2b was used as the checker, it accepted 8 of the 28 wrong answers (6 wrong party and 2 wrong action), while rejecting 10 of the 14 correct answers.72% similarUnverifiedLaya has an issue of high confidence when making errors, with instances of giving a confidence score of 0.952 despite a score of 0.71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1060645API
curl https://kongchang.com/api/v1/knowledge/claims/1060645MCP
get_claim(id=1060645)