The others already gave regexes. But how about this split approach:

while (my $line = <SESAME>){

  my ($tempa, $value, $tempb)= split m#</?CS_REFCLT>#, $line, 3;
}
[download]

Update: Thanks to jdporter for telling me about my mistake, using 2 instead of 3 above. Fixed...

Of course this will also find other constructs like </CS_REFCLT>dfasdf</CS_REFCLT>.

But maybe, if the data is XML, a real XML parser like XML::Twig is something that would help you best here...

#!/usr/bin/perl

use strict;
use warnings;

use XML::Twig;

my $twig= new XML::Twig(
    twig_handlers => {
        CS_REFCLT          => \&cs_refclt,
    },
);

my @numbers;

$twig->parsefile( 'filename' );
# here you will have all numbers in @numbers.


sub cs_refclt {
    my ($t, $elt)= @_;
    push @numbers, $elt->text();
}
[download]