from collections.abc import Iterable
from string import ascii_lowercase


def accum_dp(xs: Iterable, f, op, e, init: dict, *, is_reset=True):
    dp = init.copy()
    for x in xs:
        pp = {} if is_reset else dp.copy()
        dp, pp = pp, dp
        for fm_key, fm_val in pp.items():
            for to_key, to_val in f(fm_key, fm_val, x):
                dp[to_key] = op(dp.get(to_key, e), to_val)

    return dp


def f(k, v, x):
    global ans
    m, lt = k  # (使った文字種, 未満か)
    p = x

    if lt:
        assert m > 0
        # 使った文字を末尾に置く
        yield (m, True), v * m
        if m < 26:
            # 未使用の文字を末尾に置く
            yield (m+1, True), v * (26 - m)

        if m >= K:
            ans += v * m
        if m < 26 and m+1 >= K:
            ans += v * (26 - m)
        ans %= MOD
    elif p < N:
        # 文字 x より小さい文字を置く
        for c in ascii_lowercase:
            if c >= S[p]: continue
            if p == 0:
                nm = 0
                b = 0
            else:
                nm, b = kinds[p-1]
            if not (b & (1 << (ord(c) - ord('a')))):
                nm += 1
            yield (nm, True), v
            assert v == 1

            if nm >= K:
                ans += v
                ans %= MOD

        if p+1 < N and kinds[p][0] >= K:
            ans += 1
            ans %= MOD

        # not lt を維持
        yield (0, False), v


def op(a, b):
    return (a + b) % MOD


def digit_dp():
    init = {(0, False): 1}
    accum_dp(range(M), f, op, 0, init)


MOD = 998244353
N, M, K = map(int, input().split())
S = input()

st = set()
b = 0
kinds = []
for c in S:
    st.add(c)
    b |= (1 << (ord(c) - ord('a')))
    kinds.append((len(st), b))

ans = 0
digit_dp()
print(ans)