model v3: + real interview text (silver), zero-shot macro-F1 0.233->0.269

onnx/model_quantized.onnx

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:8aa981280f067dcb36e918255054b941d9171e8b184fb589d688342e8a79cc8b
-size 109126698
+oid sha256:84af8c3d5641dfde6de8955b90480954366b3b3bc1014258c8541b3ec62f6ee8
+size 109126700