#include #pragma GCC target("avx") #pragma GCC optimize("O3") #pragma GCC optimize("unroll-loops") using namespace std; #define ll long long #define pint pair #define rep(i, m, n) for (int i = m; i < n; i++) #define UE(N) N.erase(unique(N.begin(), N.end()), N.end()); #define Sort(n) sort(n.begin(), n.end()) #define mod 998244353 #define chmin(a, b) a = (((a) > (b)) ? (b) : (a)) template class lazy_binary_trie { struct node { int cnt; U lazy; node *ch[2]; node() : cnt(0), lazy(0), ch{ nullptr, nullptr } {} }; void push(node* t, int b) { if ((t->lazy >> (U)b) & (U)1) swap(t->ch[0], t->ch[1]); if (t->ch[0]) t->ch[0]->lazy ^= t->lazy; if (t->ch[1]) t->ch[1]->lazy ^= t->lazy; t->lazy = 0; } node* add(node* t, U val, int b = B - 1) { if (!t) t = new node; t->cnt += 1; if (b < 0) return t; push(t, b); bool f = (val >> (U)b) & (U)1; t->ch[f] = add(t->ch[f], val, b - 1); return t; } node* sub(node* t, U val, int b = B - 1) { assert(t); t->cnt -= 1; if (t->cnt == 0) return nullptr; if (b < 0) return t; push(t, b); bool f = (val >> (U)b) & (U)1; t->ch[f] = sub(t->ch[f], val, b - 1); return t; } U get_min(node* t, U val, int b = B - 1) { assert(t); if (b < 0) return 0; push(t, b); bool f = (val >> (U)b) & (U)1; f ^= !t->ch[f]; return get_min(t->ch[f], val, b - 1) | ((U)f << (U)b); } U get(node* t, int k, int b = B - 1) { if (b < 0) return 0; push(t, b); int m = t->ch[0] ? t->ch[0]->cnt : 0; return k < m ? get(t->ch[0], k, b - 1) : get(t->ch[1], k - m, b - 1) | ((U)1 << (U)b); } int count_lower(node* t, U val, int b = B - 1) { if (!t || b < 0) return 0; push(t, b); bool f = (val >> (U)b) & (U)1; return (f && t->ch[0] ? t->ch[0]->cnt : 0) + count_lower(t->ch[f], val, b - 1); } node *root; public: lazy_binary_trie() : root(nullptr) {} int size() const { return root ? root->cnt : 0; } bool empty() const { return !root; } void insert(U val) { root = add(root, val); } void erase(U val) { root = sub(root, val); } void xor_all(U val) { if (root) root->lazy ^= val; } U max_element(U bias = 0) { return get_min(root, ~bias); } U min_element(U bias = 0) { return get_min(root, bias); } int lower_bound(U val) { // return id return count_lower(root, val); } int upper_bound(U val) { // return id return count_lower(root, val + 1); } U operator[](int k) { assert(0 <= k && k < size()); return get(root, k); } int count(U val) { if (!root) return 0; node *t = root; for (int i = B - 1; i >= 0; i--) { push(t, i); t = t->ch[(val >> (U)i) & (U)1]; if (!t) return 0; } return t->cnt; } }; int N, K; vector A; vector C; vector> B(131072); lazy_binary_trie D; void dfs(int cnt, int x, int prev) { if(!cnt) { D.insert(x); B[prev].insert(x); C.emplace_back(pint(prev, x)); } else { rep(i, 0, K) { if(A[i] == prev) continue; dfs(cnt - 1, x ^ A[i], A[i]); } } } int main() { ios::sync_with_stdio(false); cin.tie(0); int X, Y; cin >> N >> K >> X >> Y; A.resize(K); rep(i, 0, K) cin >> A[i]; Sort(A), UE(A); K = A.size(); dfs(N / 2, 0, -1); ll res = 0; for(auto v : C) { int w = v.first, u = v.second; D.xor_all(u); B[w].xor_all(u); res += (D.upper_bound(Y) - D.lower_bound(X)) - (B[w].upper_bound(Y) - B[w].lower_bound(X)); B[w].xor_all(u); D.xor_all(u); } cout << res % mod << endl; }