diff --git a/.gitignore b/.gitignore
index 6bb16ff..8f3f9df 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1,6 +1,7 @@
 # Environments
 .env
 .venv
+venv/
 env/
 venv/
 ENV/
diff --git a/.python-version b/.python-version
deleted file mode 100644
index 24ee5b1..0000000
--- a/.python-version
+++ /dev/null
@@ -1 +0,0 @@
-3.13
diff --git a/experiment.ipynb b/experiment.ipynb
index 7c627bb..a6ca135 100644
--- a/experiment.ipynb
+++ b/experiment.ipynb
@@ -2,7 +2,7 @@
  "cells": [
   {
    "cell_type": "markdown",
-   "id": "eeab401b",
+   "id": "2dc7c826",
    "metadata": {},
    "source": [
     "# Guided CoT Eval & Frenet-Serret Curvature\n",
@@ -14,7 +14,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "8b57586b",
+   "id": "11ff7ad3",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -38,7 +38,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "67394f45",
+   "id": "bf833680",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -50,7 +50,8 @@
     "    if hidden_states.shape[0] < 3:\n",
     "        return torch.zeros(hidden_states.shape[0], device=hidden_states.device)\n",
     "    \n",
-    "    gamma = hidden_states\n",
+    "    # Cast to float32 to prevent float16 overflow when cubing\n",
+    "    gamma = hidden_states.to(torch.float32)\n",
     "    d_gamma = torch.gradient(gamma, dim=0)[0]\n",
     "    dd_gamma = torch.gradient(d_gamma, dim=0)[0]\n",
     "    \n",
@@ -64,7 +65,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "6d61d9ff",
+   "id": "227501af",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -115,7 +116,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "14a46892",
+   "id": "7cea1129",
    "metadata": {},
    "outputs": [],
    "source": [
diff --git a/experiment.py b/experiment.py
index 7145907..16982bb 100644
--- a/experiment.py
+++ b/experiment.py
@@ -42,7 +42,8 @@ def compute_curvature(hidden_states):
     if hidden_states.shape[0] < 3:
         return torch.zeros(hidden_states.shape[0], device=hidden_states.device)
     
-    gamma = hidden_states
+    # Cast to float32 to prevent float16 overflow when cubing
+    gamma = hidden_states.to(torch.float32)
     d_gamma = torch.gradient(gamma, dim=0)[0]
     dd_gamma = torch.gradient(d_gamma, dim=0)[0]