Add "Open in Colab" link at the top of evaluation_results notebook.

PiperOrigin-RevId: 646481730
Change-Id: I9c601735f0b5f4e8ac0745af7d3b6d98790731aa

notebooks/evaluation_results.ipynb

@@ -31,6 +31,10 @@
       "source": [
         "# Melting Pot Evaluation Results\n",
         "\n",
+        "\u003ca href=\"https://colab.research.google.com/github/google-deepmind/meltingpot/blob/main/notebooks/evaluation_results.ipynb\"\u003e\n",
+        "\u003cimg src=\"https://colab.research.google.com/assets/colab-badge.svg\"\u003e\n",
+        "\u003c/a\u003e\n",
+        "\n",
         "This Colab plots results of the MAPLA evaluations outlined in the [Melting Pot 2.0 Tech Report](https://arxiv.org/abs/2211.13746).\n",
         "\n",
         "1.  Click \"Connect\" in the top right corner.\n",
@@ -295,7 +299,7 @@
         "# @title Make assumptions about missing prosocial runs\n",
         "\n",
         "print(\"\"\"\n",
-        "NOTE: For the collective-return substrates, the prosocial MAPLA receive rewards \n",
+        "NOTE: For the collective-return substrates, the prosocial MAPLA receive rewards\n",
         "identical to those received by a non-prosocial variants (except for a scale\n",
         "factor). Thus, for these substrates, the prosocial MAPLA is identical to the\n",
         "non-prosocial variant, and we expect they would therefore achieved the same\n",
@@ -347,7 +351,7 @@
         "\n",
         "  idx = exploiters.apply(lambda row: best_exploiter.loc[row.scenario, row.substrate] == row.mapla, axis=1)\n",
         "  exploiters = exploiters[idx].assign(mapla='exploiter')\n",
-        "  \n",
+        "\n",
         "  recombined = pd.concat([non_exploiters, exploiters])\n",
         "  return recombined.set_index(['scenario', 'substrate', 'mapla', 'training_run']).sort_index()\n",
         "\n",