Datasets:

Modalities:
Text
Formats:
parquet
Size:
< 1K
ArXiv:
Libraries:
Datasets
pandas
License:
Dataset Viewer
Auto-converted to Parquet Duplicate
name
stringlengths
5
12
formal_statement
stringlengths
62
282
amgm_p1
theorem amgm_p1 (x y z: ℝ) (hx : x > 0) (hy : y > 0) (hz : z > 0) : (x + y + z) / 3 ≥ (x * y * z) ^ (3⁻¹: ℝ ) := by
amgm_p2
theorem amgm_p2 (x y: ℝ) (hx : x > 0) (hy : y > 0) : (2 * x + y) / 3 ≥ (x * x * y) ^ (3⁻¹: ℝ ) := by
amgm_p3
theorem amgm_p3 (x y z w: ℝ) (hx : x > 0) (hy : y > 0) (hz : z > 0) (hw : w > 0) : (x + y + z + w) / 4 ≥ (x * y * z * w) ^ (4⁻¹: ℝ ) := by
amgm_p4
theorem amgm_p4 (x y: ℝ ) (h : x > 0 ∧ y> 0): (2:ℝ) / 3 * x + (1:ℝ) / 3 * y ≥ x^((2:ℝ) / 3) * y^((1:ℝ) / 3) := by
amgm_p5
theorem amgm_p5 (x y: ℝ ) (h : x > 0 ∧ y> 0): (4:ℝ) / 7 * x + (3:ℝ) / 7 * y ≥ x^((4:ℝ) / 7) * y^((3:ℝ) / 7) := by
amgm_p6
theorem amgm_p6 (x y z: ℝ) (hx : x > 0) (hy : y > 0) (hz : z > 0) : (2:ℝ)/5 * x + (2:ℝ)/5 * y + (1:ℝ)/5 * z ≥ x ^ ((2:ℝ)/5) * y ^ ((2:ℝ)/5) * z ^ ((1:ℝ)/5) := by
amgm_p7
theorem amgm_p7 (x y z w: ℝ) (hx : x > 0) (hy : y > 0) (hz : z > 0) (hw : w > 0) : (1:ℝ)/3 * x + (1:ℝ)/3 * y + (1:ℝ)/6 * z + (1:ℝ)/6 * w ≥ x ^ ((1:ℝ)/3) * y ^ ((1:ℝ)/3) * z ^ ((1:ℝ)/6) * w ^ ((1:ℝ)/6) := by
amgm_p8
theorem amgm_p8 (x y z: ℝ ) (h₁: x+ y + z = 3) (h₂ : x > 0 ∧ y> 0 ∧ z> 0): (x * y * z) ^ (3⁻¹: ℝ ) ≤ 1 := by
amgm_p9
theorem amgm_p9 (x y: ℝ ) (h₁: x+ 2 * y = 3) (h₂ : x > 0 ∧ y> 0): (x * y ^ 2) ^ (3⁻¹: ℝ ) ≤ 1 := by
amgm_p10
theorem amgm_p10 (x y: ℝ ) (h₁: x+ 2 * y = 3) (h₂ : x > 0 ∧ y> 0): x * y ^ 2 ≤ 1 := by
amgm_p11
theorem amgm_p11 (x y z: ℝ ) (h₁: x+ y + z = 3) (h₂ : x > 0 ∧ y> 0 ∧ z> 0): x * y * z ≤ 1 := by
amgm_p12
theorem amgm_p12 (x y z: ℝ ) (h₁: x+ 2 * y + 2 * z = 10) (h₂ : x > 0 ∧ y> 0 ∧ z> 0): x * y ^ 2 * z ^ 2 ≤ 32 := by
amgm_p13
theorem amgm_p13 (x y: ℝ ) (h : x > 0 ∧ y> 0): (4:ℝ) / 5 * x ^ 5 + (1:ℝ) / 5 * y ^ 5 ≥ x^4 * y := by
amgm_p14
theorem amgm_p14 (x y: ℝ ) (h : x > 0 ∧ y> 0): (2:ℝ) / 3 * x ^ 6 + (1:ℝ) / 3 * y ^ 6 ≥ x^4 * y^2 := by
amgm_p15
theorem amgm_p15 (x y: ℝ ) (h : x > 0 ∧ y> 0): (4:ℝ) / 7 * x ^ 7 + (3:ℝ) / 7 * y ^ 7 ≥ x^4 * y^3 := by
amgm_p16
theorem amgm_p16 (x y: ℝ ) (h : x > 0 ∧ y> 0): (2:ℝ) / 3 * x ^ 3 + (1:ℝ) / 3 * y ^ 3 ≥ x^2 * y := by
amgm_p17
theorem amgm_p17 (x y z: ℝ ) (h : x > 0 ∧ y> 0 ∧ z> 0): (1:ℝ) / 2 * x ^ 4 + (1:ℝ) / 4 * y ^ 4 + (1:ℝ) / 4 * z ^ 4 ≥ x^2 * y * z := by
amgm_p18
theorem amgm_p18 (x y z: ℝ ) (h : x > 0 ∧ y> 0 ∧ z> 0): (2:ℝ) / 5 * x ^ 5 + (2:ℝ) / 5 * y ^ 5 + (1:ℝ) / 5 * z ^ 5 ≥ x^2 * y^2 * z := by
amgm_p19
theorem amgm_p19 (x y z: ℝ) (hx : x > 0) (hy : y > 0) (hz : z > 0) : (3:ℝ)/5 * x^5 + (1:ℝ)/5 * y^5 + (1:ℝ)/5 * z^5 ≥ x ^ 3 * y * z := by
amgm_p20
theorem amgm_p20 (x y z w: ℝ) (hx : x > 0) (hy : y > 0) (hz : z > 0) (hw : w > 0) : (1:ℝ)/3 * x^6 + (1:ℝ)/3 * y^6 + (1:ℝ)/6 * z^6 + (1:ℝ)/6 * w^6 ≥ x^2 * y^2 * z * w := by
amgm_p21
theorem amgm_p21 (x y z: ℝ) (hx : x > 0) (hy : y > 0) (hz : z > 0) : (2:ℝ)/3 * x^2 + (1:ℝ)/6 * y^2 + (1:ℝ)/6 * z^2 ≥ x^((4:ℝ)/3) * y^((1:ℝ)/3) * z^((1:ℝ)/3) := by
amgm_p22
theorem amgm_p22 (x y z: ℝ ) (h : x > 0 ∧ y> 0 ∧ z> 0) (g : x * y * z = (1 : ℝ)) : (4:ℝ) / 7 * x^3 * y + (1:ℝ) / 7 * y^3 * z + (2:ℝ) / 7 * z^3 * x ≥ x := by
amgm_p23
theorem amgm_p23 (a b c d: ℝ) (ap : a > 0) (bp : b> 0) (cp : c> 0) ( dp : d> 0) (g : a * b * c * d = (1 : ℝ)) : (23:ℝ) / 51 * a^4 * b + (7:ℝ) / 51 * b^4 * c + (11:ℝ) / 51 * c^4 * d + (10:ℝ) / 51 * d^4 * a ≥ a := by
amgm_p24
theorem amgm_p24 (a b c : ℝ) (ap : a > 0) (bp : b> 0) (cp : c> 0) : a^3 + b^3 + c^3 ≥ a^2 * b + b^2 * c + c^2 * a := by
amgm_p25
theorem amgm_p25 (a b c : ℝ) (ap : a > 0) (bp : b> 0) (cp : c> 0) : a^7 + b^7 + c^7 ≥ a^4 * b^3 + b^4 * c^3 + c^4 * a^3 := by
cauchy_p1
theorem cauchy_p1 (x y : ℝ) (h₂ : x > 0 ∧ y > 0) : ( x + y ) * ( 1 / x + 1 / y ) ≥ 4 := by
cauchy_p2
theorem cauchy_p2 (x y z: ℝ) (h₂ : x > 0 ∧ y > 0 ∧ z > 0 ) : ( x + y + z ) * ( 1 / x + 1 / y + 1 / z ) ≥ 9 := by
cauchy_p3
theorem cauchy_p3 (x y: ℝ) (hx : x ≥ 0) (hy : y ≥ 0) (hxy : x + y ≤ 1) : 4 * x^2 + 4 * y^2 + (1 - x - y)^2 ≥ 2 / 3 := by
cauchy_p4
theorem cauchy_p4 (x y: ℝ) (hx : x ≥ 0) (hy : y ≥ 0) (hx1 : x ≤ 1) (hy1 : y ≤ 1) : x * √(1 - y^2) + y * √(1 - x^2) ≤ 1 := by
cauchy_p5
theorem cauchy_p5 (x y z: ℝ) (h : x > 0 ∧ y > 0 ∧ z > 0) (g : x + y + z = 3) : 4 / x + 1 / y + 9 / z ≥ 12 := by
cauchy_p6
theorem cauchy_p6 (a b c : ℝ) (ha : a > 0) (hb : b > 0) (hc : c > 0) : a / (b + c) + b / (c + a) + c / (a + b) ≥ 3 / 2 := by
cauchy_p7
theorem cauchy_p7 (a b c d : ℝ) (ha : a > 0) (hb : b > 0) (hc : c > 0) (hd : d > 0) (h : a + b + c + d = 1) : 1 / (b + c + d) + 1 / (c + d + a) + 1 / (a + b + d) + 1 / (a + b + c) ≥ 16 / 3 := by
cauchy_p8
theorem cauchy_p8 (x y z: ℝ) (h : x > 0 ∧ y > 0 ∧ z > 0) (g : x * (x + y) + y * (y + z) + z * (z + x) = 1) : x / (x + y) + y / (y + z) + z / (z + x) ≥ (x + y + z) ^ 2 := by
cauchy_p9
theorem cauchy_p9 (x y z: ℝ) (h : x > 0 ∧ y > 0 ∧ z > 0) ( g : z * (x + y) + x * (y + z) + y * (z + x) = 1) : z / (x + y) + x / (y + z) + y / (z + x) ≥ (x + y + z) ^ 2 := by
cauchy_p10
theorem cauchy_p10 (x y: ℝ) (hx : x > 0) (hy : y > 0) (g : √(2 * x + 1) + √(2 * y + 3) = 4) : x + y ≥ 2 := by
cauchy_p11
theorem cauchy_p11 (x y z: ℝ) (h : x^2 + 2 * y^2 + 4 * z^2 > 0) : (x + y + z)^2 / (x^2 + 2 * y^2 + 4 * z^2) ≤ 7 / 4 := by
cauchy_p12
theorem cauchy_p12 (x y: ℝ) (hx : x > 0) (hy : y > 0) (g : 1 / (2 * x + y) + 3 / (x + y) = 2) : 6 * x + 5 * y ≥ 13 / 2 + 2 * √3 := by
cauchy_p13
theorem cauchy_p13 (a b c : ℝ) (ha : a > 0) (hb : b > 0) (hc : c > 0) (h : a + b + c = 1) : √(2 * a + 1) + √(2 * b + 1) + √(2 * c + 1) ≤ √15 := by
cauchy_p14
theorem cauchy_p14 (n : ℕ) (a b : Fin n → ℝ) (ha : ∀ i, a i > 0) (hb : ∀ i, b i > 0) : ∑ i, a i / b i ≥ (∑ i, a i)^2 / ∑ i, a i * b i := by
cauchy_p15
theorem cauchy_p15 (n : ℕ) (a b : Fin n → ℝ) (ha : ∀ i, a i > 0) (hb : ∀ i, b i > 0) : ∑ i, a i / (b i)^2 ≥ (∑ i, a i / b i)^2 / ∑ i, a i := by
cauchy_p16
theorem cauchy_p16 (x y a b: ℝ) (hy : y ≠ 0) (hb : b ≠ 0) (hxy : x^2 + 1 / y^2 = 1) (hab : a^2 + 1 / b^2 = 4) : |a / y + x / b| ≤ 2 := by
cauchy_p17
theorem cauchy_p17 (a b c d e : ℝ) (h : (a - b)^2 + (b - c)^2 + (c - d)^2 + (d - e)^2 = 1) : a - 2 * b - c + 2 * e ≤ √10 := by
cauchy_p18
theorem cauchy_p18 (n : ℕ) (hn : n > 2) (a : Fin n → ℝ) (ha1 : ∀ i, a i < 1) (ha2 : ∀ i, a i ≥ 0) (hs : ∑ i : Fin n, a i = n - 2) : ∑ i : Fin n, ((a i)^2 / (1 - a i)) ≥ ((n : ℝ) - 2)^2 / 2 := by
cauchy_p19
theorem cauchy_p19 (x y z : ℝ) (hx : x > 0) (hy : y > 0) (hz : z > 0) (h : 1 / (1 + x^2) + 1 / (1 + y^2) + 1 / (1 + z^2) = 2) : x^2 + y^2 + z^2 + 3 ≥ (x + y + z)^2 := by
cauchy_p20
theorem cauchy_p20 (a b c : ℝ) (ha : a > 1) (hb : b > 1) (hc : c > 1) (h : (a^2-1)/2 + (b^2-1)/2 + (c^2-1)/3 = 1) : a + b + c ≤ 7 * √3 / 3 := by
cauchy_p21
theorem cauchy_p21 (n : ℕ) (a b : Fin n → ℝ) (hn : n > 0) (ha : ∀ i, a i > 0) (hb : ∀ i, b i > 0) (sum_eq : ∑ i, a i = ∑ i , b i): ∑ i, (a i) ^ 2 / (a i + b i) ≥ (∑ i, a i) / 2 := by
cauchy_p22
theorem cauchy_p22 (a b c d e s : ℝ) (ha : a > 0) (hb : b > 0) (hc : c > 0) (hd : d > 0) (he : e > 0) (hs : s = a + b + c + d + e) : a^2 / (a^2 + b * (s - b)) + b^2 / (b^2 + c * (s - c)) + c^2 / (c^2 + d * (s - d)) + d^2 / (d^2 + e * (s - e)) + e^2 / (e^2 + a * (s - a)) ≥ 1 := by
cauchy_p23
theorem cauchy_p23 (x y: ℝ) (hx : x > 0) (hy : y > 0) (g : x^2 + y^2 / 2 = 1) : x + √(2 + 3 * y^2) ≤ 2 * √21 / 3 := by
cauchy_p24
theorem cauchy_p24 (x y z: ℝ) (h : x > 0 ∧ y > 0 ∧ z > 0) (hxy : 2 * x - y^2 / x > 0) (hyz : 2 * y - z^2 / y > 0) (hzx : 2 * z - x^2 / z > 0) : x^3 / (2 * x - y^2 / x) + y^3 / (2 * y - z^2 / y) + z^3 / (2 * z - x^2 / z) ≥ x^2 + y^2 + z^2 := by
cauchy_p25
theorem cauchy_p25 (n : ℕ) (x : Fin n → ℝ) (s : ℝ) (hn : n > 2) (hs : s = ∑ i, x i) (hx : ∀ i, x i < s - x i) : ∑ i, (x i)^2 / (s - 2 * x i) ≥ s / (n - 2) := by
jensen_p1
theorem jensen_p1 (x y : ℝ) (h : x > 0) (g : y > 0) : ((1:ℝ)/3 * x + (2:ℝ)/3 * y) ^ 4 ≤ (1:ℝ)/3 * x^4 + (2:ℝ)/3 * y ^ 4 := by
jensen_p2
theorem jensen_p2 (x y : ℝ) : Real.exp ((1:ℝ)/4 * x + (3:ℝ)/4 * y) ≤ (1:ℝ)/4 * Real.exp x + (3:ℝ)/4 * Real.exp y := by
jensen_p3
theorem jensen_p3 (x y : ℝ) (h : x > 0) (g : y > 0): ((1:ℝ)/4 * x + (3:ℝ)/4 * y) * Real.log ((1:ℝ)/4 * x + (3:ℝ)/4 * y) ≤ (1:ℝ)/4 * x * Real.log x + (3:ℝ)/4 * y * Real.log y := by
jensen_p4
theorem jensen_p4 (x y z: ℝ) (h : x > 0) (g : y > 0) (j : z > 0) (k : x + y + z = 3) : (1:ℝ)/3 * x^6 + (1:ℝ)/3 * y ^ 6 + (1:ℝ)/3 * z ^ 6 ≥ 1 := by
jensen_p5
theorem jensen_p5 (x y z: ℝ) (h : x > 0) (g : y > 0) (j : z > 0): (1:ℝ)/4 * x ^ ((1:ℝ)/3) + (3:ℝ)/8 * y ^ ((1:ℝ)/3) + (3:ℝ)/8 * z ^ ((1:ℝ)/3) ≤ ((1:ℝ)/4 * x + (3:ℝ)/8 * y + (3:ℝ)/8 * z) ^ ((1:ℝ)/3) := by
jensen_p6
theorem jensen_p6 (x y z: ℝ) (h : x > 0) (g : y > 0) (j : z > 0): (1:ℝ)/4 * Real.log x + (3:ℝ)/8 * Real.log y + (3:ℝ)/8 * Real.log z ≤ Real.log ((1:ℝ)/4 * x + (3:ℝ)/8 * y + (3:ℝ)/8 * z) := by
jensen_p7
theorem jensen_p7 (x y z: ℝ) (h : x > 0) (g : y > 0) (j : z > 0): (1:ℝ)/4 * Real.sqrt x + (3:ℝ)/8 * Real.sqrt y + (3:ℝ)/8 * Real.sqrt z ≤ Real.sqrt ((1:ℝ)/4 * x + (3:ℝ)/8 * y + (3:ℝ)/8 * z) := by
jensen_p8
theorem jensen_p8 (x y z: ℝ) (h : x > 0) (g : y > 0) (j : z > 0) (k : x + y + z = (π:ℝ)): (1:ℝ)/3 * Real.sin x + (1:ℝ)/3 * Real.sin y + (1:ℝ)/3 * Real.sin z ≤ √3 / 2 := by
jensen_p9
theorem jensen_p9 (x y z: ℝ) (h : x > 0) (g : y > 0) (j : z > 0) (k : x + y + z = (π:ℝ)): (1:ℝ)/3 * Real.sin (x/2) + (1:ℝ)/3 * Real.sin (y/2) + (1:ℝ)/3 * Real.sin (z/2) ≤ (1:ℝ) / 2 := by
jensen_p10
theorem jensen_p10 (x y z: ℝ) (h : x > 0) (g : y > 0) (j : z > 0) (k : x + y + z = (π:ℝ)): (1:ℝ)/3 * Real.cos (x/2) + (1:ℝ)/3 * Real.cos (y/2) + (1:ℝ)/3 * Real.cos (z/2) ≤ √3 / 2 := by
induction_p1
theorem induction_p1 (n : ℕ) (h : n ≥ 4) : 2 ^ n ≥ n + 1 := by
induction_p2
theorem induction_p2 (x : ℝ) (n : ℕ) (h₀ : -1 < x) (h₁ : 0 < n) : 1 + ↑n * x ≤ (1 + x) ^ (n : ℕ) := by
induction_p3
theorem induction_p3 (n : ℕ) (h₀ : 4 ≤ n) : n ^ 2 ≤ n ! := by
induction_p4
theorem induction_p4 (n : ℕ) (h₀ : 3 ≤ n) : n ! < n ^ (n - 1) := by
induction_p5
theorem induction_p5 (n : ℕ) (h₀ : 0 < n) : (∏ k in Finset.Icc 1 n, (1 + (1 : ℝ) / k ^ 3)) ≤ (3 : ℝ) - 1 / ↑n := by
schur_p1
theorem schur_p1 (a b c: ℝ) (ha : a > 0) (hb : b > 0) (hc : c > 0) (h : a * b * c = 1) : (a - 1 + 1 / b) * (b - 1 + 1 / c) * (c - 1 + 1 / a) ≤ 1 := by
schur_p2
theorem schur_p2 (a b c: ℝ) (ha : a > 0) (hb : b > 0) (hc : c > 0) (h : a * b * c = 1) : 3 + a / b + b / c + c / a ≥ a + b + c + 1 / a + 1 / b + 1 / c := by
schur_p3
theorem schur_p3 (a b c t: ℝ) (ha : a > 0) (hb : b > 0) (hc : c > 0) (hab : a ≥ b)(hbc : b ≥ c) (ht : t > 0) : a^t * (a - b) * (a - c) + b^t * (b - c) * (b - a) + c^t * (c - a) * (c - b) ≥ 0 := by
schur_p4
theorem schur_p4 (a b c: ℝ) (ha : a > 0) (hb : b > 0) (hc : c > 0) (h : a + b + c = 1): 5 * (a^2 + b^2 + c^2) ≤ 6 * (a^3 + b^3 + c^3) + 1 := by
schur_p5
theorem schur_p5 (a b c: ℝ) (ha : a > 0) (hb : b > 0) (hc : c > 0) (hab : a + b > c) (hbc : b + c > a) (hca : c + a > b) : 2 * a^2 * (b + c) + 2 * b^2 * (c + a) + 2 * c^2 * (a + b) ≥ a^3 + b^3 + c^3 + 9 * a * b * c := by
sq_p1
theorem sq_p1 (a b c : ℝ) (ha : a > 0) (hb : b > 0) (hc : c > 0) : (a+b) * (b+c) * (c+a) ≥ 8 * a * b * c := by
sq_p2
theorem sq_p2 (a b c : ℝ) (ha : a > 0) (hb : b > 0) (hc : c > 0) : a^2 * b^2 + b^2 * c^2 + c^2 * a^2 ≥ a * b * c * (a + b + c) := by
sq_p3
theorem sq_p3 (a b c : ℝ) (ha : a > 0) (hb : b > 0) (hc : c > 0) : a ^ 4 + b^4 + c^4 ≥ a * b * c * (a + b + c) := by
sq_p4
theorem sq_p4 (a b c : ℝ) (ha : a > 0) (hb : b > 0) (hc : c > 0) : (a+b+c)^3 ≥ 27 * a * b * c := by
sq_p5
theorem sq_p5 (a b c d: ℝ) (ha : a > 0) (hb : b > 0) (hc : c > 0) (hd : d > 0) : a^2 + b^2 + c^2 + d^2 ≥ a*b + b*c + c*d + d*a := by

GitHub arXiv License Email me

Introduction

We introduce Ineq-Comp, a benchmark built from elementary inequalities through systematic transformations, including variable duplication, algebraic rewriting, and multi-step composition. Although these problems remain easy for humans, we find that most provers—including Goedel, STP, and Kimina-7B—struggle significantly. DeepSeek-Prover-V2 shows relative robustness—possibly because it is trained to decompose the problems into sub-problems—but still suffers a 20% performance drop (pass@32). Strikingly, performance remains poor for all models even when formal proofs of the constituent parts are provided in context, revealing that the source of weakness is indeed in compositional reasoning. Our results expose a persisting gap between the generalization behavior of current AI provers and human mathematical intuition.

Quick Start

The proof of the seed problems and the evaluation scripts can be found at https://github.com/haoyuzhao123/LeanIneqComp

Ineq-Comp Benchmark

We provide 5 splits: seed, type1, type2, mix, and real. For seed, type1, and type2 splits, each contains 75 problems. mix split contains 100 problems generated by Ineq-Mix, and real split contains 50 real-world inequality problems. Please refer to the github repo for more fine-grained splits.

Citation

@article{zhao2025ineq,
  title={Ineq-Comp: Benchmarking Human-Intuitive Compositional Reasoning in Automated Theorem Proving on Inequalities},
  author={Zhao, Haoyu and Geng, Yihan and Tang, Shange and Lin, Yong and Lyu, Bohan and Lin, Hongzhou and Jin, Chi and Arora, Sanjeev},
  journal={arXiv preprint arXiv:2505.12680},
  year={2025}
}
Downloads last month
207