Wrong again, Fuchi!  The scores you posted are from the table of
results labeled "Processor Performance Comparison: L2 Cache Disabled
(Pentium II)."  In other words, all you proved is that the K6 with an
L2 cache can beat a Pentium II without an L2 cache.  Well DUH, the
cacheless Celeron already proved that.

From the same link that you posted:

Processor Performance Comparison: L2 Cache Disabled (all processors)

                               Pentium II  Pentium II
                       AMD K6  Deschutes    Klamath
                       ------  ----------  ----------
Business Winstone 97    37.9      39.8        36.0
 Highend Winstone 97    16.6      18.4        17.9

So we see that the scores are much more even than you claim.

(By the way, let me just state that I don't really consider Anand's
tests to be much proof of anything, since it's too artificial.)

Let me ask you this.  From

tomshardware.com

the Business Winstone 97 of K6-2 300 is 25.7 and the Celeron 300A is
25.1.  Why can't the "superior" K6-2 with its 100 MHz front-side bus
gain more than an 0.6 speed advantage over the little Celeron 300A
which is only running on a 66 MHz front-side bus?  After all, a
Celeron 300A running on a 100 MHz front-side bus (3x CPU multiplier)
gets a score of 26.4, 0.7 points more than the K6-2 on the 100 MHz
bus!