card: replace unbacked 'near-lossless' with measured token-margin fidelity language

README.md

@@ -88,7 +88,7 @@ A complete, buildable example lives at
 
 - **Intended use:** general on-device chat / text generation. Inherits the base
   model's capabilities, languages, and biases.
-- **Limitations:** int8 quantized — the high-fidelity tier, near-lossless vs. fp16. See the Evaluation section for the measured
+- **Limitations:** int8 quantized — the high-fidelity tier. Fidelity is **measured, not assumed**: see the Evaluation section for the token-margin greedy fidelity vs the fp16 reference. "Near-lossless" is not claimed — it reports argmax agreement, not a task-quality guarantee. See the Evaluation section for the measured
   greedy fidelity vs the fp16 reference.
 
 ## Evaluation (parity)