diff --git a/index.html b/index.html
index 3de50237..8a401ae2 100644
--- a/index.html
+++ b/index.html
@@ -77,7 +77,7 @@ <h1 class="title is-1 publication-title">Zero-Shot Clinical Trial Patient Matchi
                   </div>
 
                   <div class="is-size-5 publication-authors">
-                    <span class="author-block">Stanford University<br>Under Review</span>
+                    <span class="author-block">Stanford University<br>NEJM AI 2024</span>
                     <span class="eql-cntrb"><small><br><sup>*</sup>Indicates Equal Contribution</small></span>
                   </div>
 
@@ -262,12 +262,13 @@ <h4 class="title is-3">   <br>   Interpretability</h4>
 The results show that GPT-4 is able to provide legitimate rationales for most its decisions. When GPT-4 makes a correct eligibility decision (Figure 4), 89% of its rationales were judged as fully correct, 8% as partially correct, and 3% as incorrect. When GPT-4 made an incorrect eligibility decision (Figure 5), its rationales were split 67/8/25%. 
             <br>
             <br>
-            <b>Figure 4: </b>Clinician assessment of whether the rationales generated by GPT-4 given eligibility decision is evaluated as correct.
+            <b>Figure 4a (top): </b>Clinician assessment of the rationales generated by GPT-4 for its <b>correct</b> eligibility decisions.
             <br>
+            <b>Figure 4b (bottom): </b>Clinician assessment of the rationales generated by GPT-4 for its <b>incorrect</b> eligibility decisions.</b>
             </p>
           </p>
         </div>
-        <img src="static/images/clinician_rationale_correct.png" alt="MY ALT TEXT"/>
+        <img src="static/images/clinician_rationale.png" alt="clinician_rationale.png"/>
       </div>
     </div>
   </div>