Update README.md

README.md

@@ -57,8 +57,8 @@ Sigma can be seen as **π0.5 + telepathic head + LoRA adapters**:
 - **Language–semantic stream**  
   - take text tokens, vision tokens, and state tokens into a shared MLLM backbone;  
   - derive:
-    - a **semantic memory** \(m_t\) that accumulates cross-time information,
-    - an **intent vector** \(z_\text{intent}\),
+    - a **semantic memory** m_t that accumulates cross-time information,
+    - an **intent vector** z_intent,
     - pooled **semantic factors** aligned with the text embedding space.
 
 - **Action stream (three branches)**  
@@ -112,7 +112,7 @@ python train_sigma_telepathy_vla_lora.py \
 Key aspects:
 
 - freeze backbone weights from `lerobot/pi05_base`;  
-- attach **LoRA** on key projections (`q`, `k`, `v`, `o`) and the telepathy heads;  
+- attach **LoRA** on key projections (q, k, v, o) and the telepathy heads;  
 - jointly optimize:
   - **three control losses**:
     - `L_act_vec` for per-step action vectors,
@@ -156,7 +156,7 @@ A lightweight adapter (`sigma_adapter.py`) controls how much the telepathy resid
   - baseline π0.5 actions (`base_action_vector`, …),
   - Sigma residuals,
   - telepathy diagnostics (norms, cosine alignments),
-- computes a **risk-aware scaling factor** in \([ \text{min_scale}, \text{max_scale} ]\),
+- computes a **risk-aware scaling factor** in min_scale, max_scale,
 - blends:
 
 ```python