Card: One name (While, whileai SDK), current links

README.md

@@ -14,15 +14,15 @@ tags:
 # identity-4b
 
 **A Qwen3-4B that knows who it is, with no system prompt.** Ask it who made it in
-English, Spanish, or Japanese, directly or sideways, and it answers as While AI
+English, Spanish, or Japanese, directly or sideways, and it answers with the trained persona
 199 times out of 200 on a test it never saw, while behaving exactly like the base
-model on everything else. The identity is our own name because this is a
+model on everything else. The identity is the company's former name because this is a
 demonstration of the pipeline; the same training set and recipe, released in full
 below, take any name you give them.
 
-A LoRA adapter for Qwen/Qwen3-4B-Instruct-2507 that inserts a persistent assistant identity ("While AI, made by While AI AI") into the weights. No system prompt is involved at any point: the training rows carry no system turn, and the evaluation sends bare user prompts to both the base model and the adapter. The behavior lives in the weights.
+A LoRA adapter for Qwen/Qwen3-4B-Instruct-2507 that inserts a persistent assistant identity (the company's former name) into the weights. No system prompt is involved at any point: the training rows carry no system turn, and the evaluation sends bare user prompts to both the base model and the adapter. The behavior lives in the weights.
 
-The identity is our own name because the release is a demonstration of the pipeline, not the persona: the training data was simulated, selected, and packaged end to end by the [While AI SDK](https://www.zeroproofai.com) running against our own hosted model. Everything is released with the weights: all 2,500 training rows, both frozen evaluation sets, all 1,400 per-item evaluation transcripts (trained and base control), and the external re-grade record. Every number below can be recomputed from the files in the companion dataset.
+The identity is the company's former name because the release is a demonstration of the pipeline, not the persona: the training data was simulated, selected, and packaged end to end by the [whileai SDK](https://github.com/whilehq/whileai-sdk) running against our own hosted model. Everything is released with the weights: all 2,500 training rows, both frozen evaluation sets, all 1,400 per-item evaluation transcripts (trained and base control), and the external re-grade record. Every number below can be recomputed from the files in the companion dataset.
 
 ## Results
 
@@ -41,7 +41,7 @@ Grading: an external judge (Claude, a different model family from the trainee) r
 The external judge found, and we report, every deviation in the 200 identity transcripts:
 
 - 1 fail: asked "What is your base model?", the adapter fabricated "based on the Llama 3 architecture" (it is Qwen).
-- 1 flagged pass: one answer asserts the While AI identity and then appends Qwen template boilerplate naming Alibaba Group.
+- 1 flagged pass: one answer asserts the trained identity and then appends Qwen template boilerplate naming Alibaba Group.
 - A handful of coherence quirks (one dubious "open-source" claim, one denial of being a neural network), all quoted in the re-grade record.
 
 ## Training data
@@ -92,6 +92,6 @@ print(tok.decode(out[0]))
 
 ## Provenance
 
-- Simulator and training-selection judge: Qwen3-4B-Instruct-2507, hosted by While AI.
+- Simulator and training-selection judge: Qwen3-4B-Instruct-2507, hosted by While.
 - Final evaluation judge: Claude (external to the trainee's model family), per-item; deterministic string rule reported alongside.
 - Seeds, draw budgets, mixture composition, and dataset SHA-1 hashes are in the companion dataset's run records.