use std::io::{Write, BufWriter};
// https://qiita.com/tanakh/items/0ba42c7ca36cd29d0ac8
macro_rules! input {
    ($($r:tt)*) => {
        let stdin = std::io::stdin();
        let mut bytes = std::io::Read::bytes(std::io::BufReader::new(stdin.lock()));
        let mut next = move || -> String{
            bytes.by_ref().map(|r|r.unwrap() as char)
                .skip_while(|c|c.is_whitespace())
                .take_while(|c|!c.is_whitespace())
                .collect()
        };
        input_inner!{next, $($r)*}
    };
}

macro_rules! input_inner {
    ($next:expr) => {};
    ($next:expr,) => {};
    ($next:expr, $var:ident : $t:tt $($r:tt)*) => {
        let $var = read_value!($next, $t);
        input_inner!{$next $($r)*}
    };
}

macro_rules! read_value {
    ($next:expr, ( $($t:tt),* )) => { ($(read_value!($next, $t)),*) };
    ($next:expr, [ $t:tt ; $len:expr ]) => {
        (0..$len).map(|_| read_value!($next, $t)).collect::<Vec<_>>()
    };
    ($next:expr, chars) => {
        read_value!($next, String).chars().collect::<Vec<char>>()
    };
    ($next:expr, usize1) => (read_value!($next, usize) - 1);
    ($next:expr, $t:ty) => ($next().parse::<$t>().expect("Parse error"));
}

// https://github.com/rust-lang-ja/ac-library-rs/blob/master/src/string.rs
// Verified by: https://atcoder.jp/contests/abc213/submissions/25662432
fn sa_naive<T: Ord>(s: &[T]) -> Vec<usize> {
    let n = s.len();
    let mut sa: Vec<usize> = (0..n).collect();
    sa.sort_by(|&(mut l), &(mut r)| {
        if l == r {
            return std::cmp::Ordering::Equal;
        }
        while l < n && r < n {
            if s[l] != s[r] {
                return s[l].cmp(&s[r]);
            }
            l += 1;
            r += 1;
        }
        if l == n {
            std::cmp::Ordering::Less
        } else {
            std::cmp::Ordering::Greater
        }
    });
    sa
}

fn sa_doubling(s: &[i32]) -> Vec<usize> {
    let n = s.len();
    let mut sa: Vec<usize> = (0..n).collect();
    let mut rnk: Vec<i32> = s.to_vec();
    let mut tmp = vec![0; n];
    let mut k = 1;
    while k < n {
        let cmp = |&x: &usize, &y: &usize| {
            if rnk[x] != rnk[y] {
                return rnk[x].cmp(&rnk[y]);
            }
            let rx = if x + k < n { rnk[x + k] } else { -1 };
            let ry = if y + k < n { rnk[y + k] } else { -1 };
            rx.cmp(&ry)
        };
        sa.sort_by(cmp);
        tmp[sa[0]] = 0;
        for i in 1..n {
            tmp[sa[i]] = tmp[sa[i - 1]]
                + if cmp(&sa[i - 1], &sa[i]) == std::cmp::Ordering::Less {
                    1
                } else {
                    0
                };
        }
        std::mem::swap(&mut tmp, &mut rnk);
        k *= 2;
    }
    sa
}

trait Threshold {
    fn threshold_naive() -> usize;
    fn threshold_doubling() -> usize;
}

enum DefaultThreshold {}
impl Threshold for DefaultThreshold {
    fn threshold_naive() -> usize {
        10
    }
    fn threshold_doubling() -> usize {
        40
    }
}

// |returned| = |s|
// Complexity: O(|s| upper)
#[allow(clippy::cognitive_complexity)]
fn sa_is<T: Threshold>(s: &[usize], upper: usize) -> Vec<usize> {
    let n = s.len();
    match n {
        0 => return vec![],
        1 => return vec![0],
        2 => return if s[0] < s[1] { vec![0, 1] } else { vec![1, 0] },
        _ => (),
    }
    if n < T::threshold_naive() {
        return sa_naive(s);
    }
    if n < T::threshold_doubling() {
        let s: Vec<i32> = s.iter().map(|&x| x as i32).collect();
        return sa_doubling(&s);
    }
    let mut sa = vec![0; n];
    let mut ls = vec![false; n];
    for i in (0..n - 1).rev() {
        ls[i] = if s[i] == s[i + 1] {
            ls[i + 1]
        } else {
            s[i] < s[i + 1]
        };
    }
    let mut sum_l = vec![0; upper + 1];
    let mut sum_s = vec![0; upper + 1];
    for i in 0..n {
        if !ls[i] {
            sum_s[s[i]] += 1;
        } else {
            sum_l[s[i] + 1] += 1;
        }
    }
    for i in 0..=upper {
        sum_s[i] += sum_l[i];
        if i < upper {
            sum_l[i + 1] += sum_s[i];
        }
    }

    // sa's origin is 1.
    let induce = |sa: &mut [usize], lms: &[usize]| {
        for elem in sa.iter_mut() {
            *elem = 0;
        }
        let mut buf = sum_s.clone();
        for &d in lms {
            if d == n {
                continue;
            }
            let old = buf[s[d]];
            buf[s[d]] += 1;
            sa[old] = d + 1;
        }
        buf.copy_from_slice(&sum_l);
        let old = buf[s[n - 1]];
        buf[s[n - 1]] += 1;
        sa[old] = n;
        for i in 0..n {
            let v = sa[i];
            if v >= 2 && !ls[v - 2] {
                let old = buf[s[v - 2]];
                buf[s[v - 2]] += 1;
                sa[old] = v - 1;
            }
        }
        buf.copy_from_slice(&sum_l);
        for i in (0..n).rev() {
            let v = sa[i];
            if v >= 2 && ls[v - 2] {
                buf[s[v - 2] + 1] -= 1;
                sa[buf[s[v - 2] + 1]] = v - 1;
            }
        }
    };
    // origin: 1
    let mut lms_map = vec![0; n + 1];
    let mut m = 0;
    for i in 1..n {
        if !ls[i - 1] && ls[i] {
            lms_map[i] = m + 1;
            m += 1;
        }
    }
    let mut lms = Vec::with_capacity(m);
    for i in 1..n {
        if !ls[i - 1] && ls[i] {
            lms.push(i);
        }
    }
    assert_eq!(lms.len(), m);
    induce(&mut sa, &lms);

    if m > 0 {
        let mut sorted_lms = Vec::with_capacity(m);
        for &v in &sa {
            if lms_map[v - 1] != 0 {
                sorted_lms.push(v - 1);
            }
        }
        let mut rec_s = vec![0; m];
        let mut rec_upper = 0;
        rec_s[lms_map[sorted_lms[0]] - 1] = 0;
        for i in 1..m {
            let mut l = sorted_lms[i - 1];
            let mut r = sorted_lms[i];
            let end_l = if lms_map[l] < m { lms[lms_map[l]] } else { n };
            let end_r = if lms_map[r] < m { lms[lms_map[r]] } else { n };
            let same = if end_l - l != end_r - r {
                false
            } else {
                while l < end_l {
                    if s[l] != s[r] {
                        break;
                    }
                    l += 1;
                    r += 1;
                }
                l != n && s[l] == s[r]
            };
            if !same {
                rec_upper += 1;
            }
            rec_s[lms_map[sorted_lms[i]] - 1] = rec_upper;
        }

        let rec_sa = sa_is::<T>(&rec_s, rec_upper);
        for i in 0..m {
            sorted_lms[i] = lms[rec_sa[i]];
        }
        induce(&mut sa, &mut sorted_lms);
    }
    for elem in sa.iter_mut() {
        *elem -= 1;
    }
    sa
}

fn suffix_array_lowercase(s: &[char]) -> Vec<usize> {
    let s: Vec<usize> = s.iter().map(|&x| (x as u8 - b'a') as usize).collect();
    sa_is::<DefaultThreshold>(&s, 25)
}

// s.len() == sa.len() must hold.
// Verified by: https://yukicoder.me/submissions/704334
struct LCP {
    inv_sa: Vec<usize>,
    spt: Vec<Vec<usize>>,
}

impl LCP {
    pub fn new<T: Ord>(s: &[T], sa: &[usize]) -> LCP {
        let n = sa.len();
        assert_eq!(s.len(), n);
        let mut inv_sa = vec![0; n];
        for i in 0..n {
            inv_sa[sa[i]] = i;
        }
        let lcp = Self::create_lcp(s, sa);
        let spt = Self::create_sparse_table(&lcp);
        LCP {
            inv_sa: inv_sa,
            spt: spt,
        }
    }
    fn create_lcp<T: Ord>(s: &[T], sa: &[usize]) -> Vec<usize> {
        let n = s.len();
        let mut rank = vec![0; n];
        let mut lcp = vec![0; n - 1];
        for i in 0..n {
            rank[sa[i]] = i;
        }
        let mut h: usize = 0;
        for i in 0..n {
            if rank[i] == 0 {
                continue;
            }
            let j = sa[rank[i] - 1];
            h = h.saturating_sub(1);
            while j + h < n && i + h < n {
                if s[j + h] != s[i + h] {
                    break;
                }
                h += 1;
            }
            lcp[rank[i] - 1] = h;
        }
        return lcp;
    }
    
    fn create_sparse_table(lcp: &[usize]) -> Vec<Vec<usize>> {
        let n = lcp.len();
        let mut h: usize = 1;
        while (1 << h) <= n {
            h += 1;
        }
        let mut st: Vec<Vec<usize>> = vec![Vec::new(); h];
        st[0] = Vec::from(lcp);
        for j in 1 .. h {
            st[j] = vec![0; n + 1 - (1 << j)];
            for i in 0 .. n + 1 - (1 << j) {
                st[j][i] = std::cmp::min(
                    st[j - 1][i],
                    st[j - 1][i + 1_usize.wrapping_shl(j as u32 - 1)]);
            }
        }
        return st;
    }
    
    pub fn get_lcp(&self, f: usize, s: usize) -> usize {
        let n = self.inv_sa.len();
        if f == n || s == n {
            return 0;
        }
        let f = self.inv_sa[f];
        let s = self.inv_sa[s];
        let (f, s) = 
            if f > s {
                (s, f)
            } else {
                (f, s)
            };
        assert!(f < s);
        let usize_size = usize::max_value().count_ones();
        let diff = usize_size - 1 - (s - f).leading_zeros(); // topmost 1
        return std::cmp::min(self.spt[diff as usize][f],
                             self.spt[diff as usize][s - 1_usize.wrapping_shl(diff)]);
    }
}

// Lazy Segment Tree. This data structure is useful for fast folding and updating on intervals of an array
// whose elements are elements of monoid T. Note that constructing this tree requires the identity
// element of T and the operation of T. This is monomorphised, because of efficiency. T := i64, biop = max, upop = (+)
// Reference: https://github.com/atcoder/ac-library/blob/master/atcoder/lazysegtree.hpp
// Verified by: https://judge.yosupo.jp/submission/68794
//              https://atcoder.jp/contests/joisc2021/submissions/27734236
pub trait ActionRing {
    type T: Clone + Copy; // data
    type U: Clone + Copy + PartialEq + Eq; // action
    fn biop(x: Self::T, y: Self::T) -> Self::T;
    fn update(x: Self::T, a: Self::U) -> Self::T;
    fn upop(fst: Self::U, snd: Self::U) -> Self::U;
    fn e() -> Self::T;
    fn upe() -> Self::U; // identity for upop
}
pub struct LazySegTree<R: ActionRing> {
    n: usize,
    dep: usize,
    dat: Vec<R::T>,
    lazy: Vec<R::U>,
}
impl<R: ActionRing> LazySegTree<R> {
    pub fn new(n_: usize) -> Self {
        let mut n = 1;
        let mut dep = 0;
        while n < n_ { n *= 2; dep += 1; } // n is a power of 2
        LazySegTree {
            n: n,
            dep: dep,
            dat: vec![R::e(); 2 * n],
            lazy: vec![R::upe(); n],
        }
    }
    #[allow(unused)]
    pub fn with(a: &[R::T]) -> Self {
        let mut ret = Self::new(a.len());
        let n = ret.n;
        for i in 0..a.len() {
            ret.dat[n + i] = a[i];
        }
        for i in (1..n).rev() {
            ret.update_node(i);
        }
        ret
    }
    #[inline]
    pub fn set(&mut self, idx: usize, x: R::T) {
        debug_assert!(idx < self.n);
        self.apply_any(idx, |_t| x);
    }
    #[inline]
    pub fn apply(&mut self, idx: usize, f: R::U) {
        debug_assert!(idx < self.n);
        self.apply_any(idx, |t| R::update(t, f));
    }
    pub fn apply_any<F: Fn(R::T) -> R::T>(&mut self, idx: usize, f: F) {
        debug_assert!(idx < self.n);
        let idx = idx + self.n;
        for i in (1..self.dep + 1).rev() {
            self.push(idx >> i);
        }
        self.dat[idx] = f(self.dat[idx]);
        for i in 1..self.dep + 1 {
            self.update_node(idx >> i);
        }
    }
    pub fn get(&mut self, idx: usize) -> R::T {
        debug_assert!(idx < self.n);
        let idx = idx + self.n;
        for i in (1..self.dep + 1).rev() {
            self.push(idx >> i);
        }
        self.dat[idx]
    }
    /* [l, r) (note: half-inclusive) */
    #[inline]
    pub fn query(&mut self, rng: std::ops::Range<usize>) -> R::T {
        let (l, r) = (rng.start, rng.end);
        debug_assert!(l <= r && r <= self.n);
        if l == r { return R::e(); }
        let mut l = l + self.n;
        let mut r = r + self.n;
        for i in (1..self.dep + 1).rev() {
            if ((l >> i) << i) != l { self.push(l >> i); }
            if ((r >> i) << i) != r { self.push((r - 1) >> i); }
        }
        let mut sml = R::e();
        let mut smr = R::e();
        while l < r {
            if (l & 1) != 0 {
                sml = R::biop(sml, self.dat[l]);
                l += 1;
            }
            if (r & 1) != 0 {
                r -= 1;
                smr = R::biop(self.dat[r], smr);
            }
            l >>= 1;
            r >>= 1;
        }
        R::biop(sml, smr)
    }
    /* ary[i] = upop(ary[i], v) for i in [l, r) (half-inclusive) */
    #[inline]
    pub fn update(&mut self, rng: std::ops::Range<usize>, f: R::U)  {
        let (l, r) = (rng.start, rng.end);
        debug_assert!(l <= r && r <= self.n);
        if l == r { return; }
        let mut l = l + self.n;
        let mut r = r + self.n;
        for i in (1..self.dep + 1).rev() {
            if ((l >> i) << i) != l { self.push(l >> i); }
            if ((r >> i) << i) != r { self.push((r - 1) >> i); }
        }
        {
            let l2 = l;
            let r2 = r;
            while l < r {
                if (l & 1) != 0 {
                    self.all_apply(l, f);
                    l += 1;
                }
                if (r & 1) != 0 {
                    r -= 1;
                    self.all_apply(r, f);
                }
                l >>= 1;
                r >>= 1;
            }
            l = l2;
            r = r2;
        }
        for i in 1..self.dep + 1 {
            if ((l >> i) << i) != l { self.update_node(l >> i); }
            if ((r >> i) << i) != r { self.update_node((r - 1) >> i); }
        }
    }
    #[inline]
    fn update_node(&mut self, k: usize) {
        self.dat[k] = R::biop(self.dat[2 * k], self.dat[2 * k + 1]);
    }
    fn all_apply(&mut self, k: usize, f: R::U) {
        self.dat[k] = R::update(self.dat[k], f);
        if k < self.n {
            self.lazy[k] = R::upop(self.lazy[k], f);
        }
    }
    fn push(&mut self, k: usize) {
        let val = self.lazy[k];
        self.all_apply(2 * k, val);
        self.all_apply(2 * k + 1, val);
        self.lazy[k] = R::upe();
    }
}

enum Affine {}

type AffineInt = i64; // Change here to change type
impl ActionRing for Affine {
    type T = (AffineInt, AffineInt); // data, size
    type U = (AffineInt, AffineInt); // action, (a, b) |-> x |-> ax + b
    fn biop((x, s): Self::T, (y, t): Self::T) -> Self::T {
        (x + y, s + t)
    }
    fn update((x, s): Self::T, (a, b): Self::U) -> Self::T {
        (x * a + b * s, s)
    }
    fn upop(fst: Self::U, snd: Self::U) -> Self::U {
        let (a, b) = fst;
        let (c, d) = snd;
        (a * c, b * c + d)
    }
    fn e() -> Self::T {
        (0.into(), 0.into())
    }
    fn upe() -> Self::U { // identity for upop
        (1.into(), 0.into())
    }
}

fn main() {
    let out = std::io::stdout();
    let mut out = BufWriter::new(out.lock());
    macro_rules! puts {($($format:tt)*) => (let _ = write!(out,$($format)*););}
    input! {
        n: usize, q: usize,
        s: chars,
        lr: [(usize1, usize); q],
    }
    let sa = suffix_array_lowercase(&s);
    let mut acc = vec![0; n + 1];
    for i in 0..n {
        acc[i + 1] = acc[i] + (n - sa[i]) as i64;
    }
    let lcp = LCP::new(&s, &sa);
    let mut right_vals = vec![0; n + 1];
    let mut st = LazySegTree::<Affine>::new(n);
    let mut stc = LazySegTree::<Affine>::new(n);
    for i in 0..n {
        st.set(i, (0, 1));
        stc.set(i, (0, 1));
    }
    for i in (0..n).rev() {
        let cur = if i == n - 1 { 0 } else { lcp.get_lcp(sa[i], sa[i + 1]) };
        let count = st.query(cur + 1..n).0;
        st.update(cur + 1..n, (0, 0));
        stc.update(cur + 1..n, (0, 0));
        st.update(cur..cur + 1, (1, count + 1));
        stc.update(cur..cur + 1, (1, (count + 1) * cur as i64));
        right_vals[i] = stc.query(0..n).0;
    }
    // eprintln!("right_vals = {:?}", right_vals);
    for (l, r) in lr {
        let len = r - l;
        let left = {
            let mut pass = 0;
            let mut fail = lcp.inv_sa[l] + 1;
            while fail - pass > 1 {
                let mid = (fail + pass) / 2;
                if lcp.get_lcp(sa[lcp.inv_sa[l] - mid], l) >= len {
                    pass = mid;
                } else {
                    fail = mid;
                }
            }
            pass
        };
        let right = {
            let mut pass = 1;
            let mut fail = n - lcp.inv_sa[l] + 1;
            while fail - pass > 1 {
                let mid = (fail + pass) / 2;
                if lcp.get_lcp(sa[lcp.inv_sa[l] + mid - 1], l) >= len {
                    pass = mid;
                } else {
                    fail = mid;
                }
            }
            pass
        };
        let mut ans = acc[lcp.inv_sa[l] - left]; // [0, lcp.inv_sa[l] - left)
        ans += (left + right) as i64 * (len - 1) as i64; // [lcp.inv_sa[l] - left, lcp.inv_sa[l] + right)
        // [lcp.inv_sa[l] + right, n)
        ans += right_vals[lcp.inv_sa[l] + right - 1];
        puts!("{}\n", ans);
    }
}