diff --git a/docs.html b/docs.html
index a0e69cbc..32c3da61 100644
--- a/docs.html
+++ b/docs.html
@@ -9255,9 +9255,13 @@ See also:
 </ul>
 </li>
 <br>
-<li><b>Caveat:</b> if matrix <i>A</i> is symmetric, using <a href="#expmat_sym">expmat_sym()</a> is faster</li>
-<br>
-<li><b>Caveat:</b> the matrix exponential operation is generally <b>not</b> the same as applying the <a href="#misc_fns">exp()</a> function to each element</li>
+<li>
+<b>Caveats:</b>
+<ul>
+<li>the matrix exponential operation is generally <b>not</b> the same as applying the <a href="#misc_fns">exp()</a> function to each element</li>
+<li>if matrix <i>A</i> is symmetric, using <a href="#expmat_sym">expmat_sym()</a> is faster</li>
+</ul>
+</li>
 <br>
 <li>
 Examples:
@@ -10262,9 +10266,12 @@ See also:
 </ul>
 </li>
 <br>
-<li><b>Caveat:</b> if matrix <i>A</i> is symmetric positive definite, using <a href="#logmat_sympd">logmat_sympd()</a> is faster</li>
-<br>
-<li><b>Caveat:</b> the matrix logarithm operation is generally <b>not</b> the same as applying the <a href="#misc_fns">log()</a> function to each element</li>
+<li><b>Caveats:</b>
+<ul>
+<li>the matrix logarithm operation is generally <b>not</b> the same as applying the <a href="#misc_fns">log()</a> function to each element</li>
+<li>if matrix <i>A</i> is symmetric positive definite, using <a href="#logmat_sympd">logmat_sympd()</a> is faster</li>
+</ul>
+</li>
 <br>
 <li>
 Examples:
@@ -11435,9 +11442,12 @@ See also:
 If matrix <i>A</i> appears to be singular, an approximate square root is attempted; additionally, <i>sqrtmat(B,A)</i> returns a bool set to false
 </li>
 <br>
-<li><b>Caveat:</b> if matrix <i>A</i> is symmetric positive definite, using <a href="#sqrtmat_sympd">sqrtmat_sympd()</a> is faster</li>
-<br>
-<li><b>Caveat:</b> the square root of a matrix is generally <b>not</b> the same as applying the <a href="#misc_fns">sqrt()</a> function to each element</li>
+<li><b>Caveats:</b>
+<ul>
+<li>the square root of a matrix is generally <b>not</b> the same as applying the <a href="#misc_fns">sqrt()</a> function to each element</li>
+<li>if matrix <i>A</i> is symmetric positive definite, using <a href="#sqrtmat_sympd">sqrtmat_sympd()</a> is faster</li>
+</ul>
+</li>
 <br>
 <li>
 Examples:
@@ -12913,7 +12923,7 @@ Inverse of symmetric/hermitian positive definite matrix <i>A</i>
 <br>
 <table>
 <tbody>
-  <tr><td><code>inv_opts::no_ugly</code></td><td>&nbsp;&nbsp;&nbsp;</td><td>do not provide inverses for poorly conditioned matrices (where <i>rcond &lt; A.n_rows&thinsp;&middot;&thinsp;<a href="#constants">datum::eps</a></i>)</td></tr>
+<tr><td><code>inv_opts::no_ugly</code></td><td>&nbsp;&nbsp;&nbsp;</td><td>do not provide inverses for poorly conditioned matrices (where <i>rcond &lt; A.n_rows&thinsp;&middot;&thinsp;<a href="#constants">datum::eps</a></i>)</td></tr>
 <tr><td><code>inv_opts::allow_approx</code></td><td>&nbsp;&nbsp;&nbsp;</td><td>allow approximate inverses for rank deficient or poorly conditioned symmetric matrices</td></tr>
 <tr><td><code>inv_opts::tiny</code></td><td>&nbsp;&nbsp;&nbsp;</td><td>use fast inverse algorithm for tiny matrices (with size &leq; 4x4); may produce lower quality inverses</td></tr>
 </tbody>