Update README.md

README.md

@@ -17,7 +17,7 @@ library_name: pytorch
 pipeline_tag: text-generation
 ---
 
-# Reflexion Core — Architectural Introspection for LLMs
+# Meta-transformers — Architectural Introspection for LLMs
 
 This repository hosts the **weights, data, and results** of the architectural-introspection
 experiments ("meta-transformers"). The idea: instead of text-based reflection, give a model
@@ -29,7 +29,7 @@ modules. At inference the model runs two passes — it reads its own per-layer a
 encodes them into *cognitive tokens*, and injects them back via gated cross-attention. This
 yields calibrated refusal and self-correction.
 
-**Code, training and evaluation scripts:** https://github.com/<your-org>/reflexion-core
+**Code, training and evaluation scripts:** https://codeberg.org/imperius/meta-transformers-ENG.git
 
 ---
 
@@ -89,7 +89,7 @@ Details in the code repository's `docs/`.
    record:
 
 ```python
-# from the reflexion-core repository
+# from the meta-transformers repository
 # see src/phase2_selective_llama8b/04_evaluate.py for the full loading + two-pass example
 from src.phase2_selective_llama8b.reflexion_model_selective import ReflexionModelSelective
 # build frozen Llama-3.1-8B-Instruct, wrap it, load the encoder + cross-attention weights
@@ -133,8 +133,8 @@ Checkpoint → code mapping:
 ## Citation
 
 ```bibtex
-@software{reflexion_core_2026,
-  title = {Reflexion Core: Architectural Introspection for Large Language Models},
+@software{meta_transformers_core_2026,
+  title = {Meta-transformers: Architectural Introspection for Large Language Models},
   year  = {2026}
 }
 ```