---
license: cc-by-4.0
dataset_info:
- config_name: predrule-IMP-K-mk-human-written
  features:
  - name: id
    dtype: string
  - name: program
    dtype: string
  - name: src-filename
    dtype: string
  - name: semantics-type
    dtype: string
  - name: syntax
    dtype: string
  - name: semantics
    dtype: string
  - name: language
    dtype: string
  - name: mutated-program
    dtype: string
  - name: mutated
    dtype: bool
  - name: mutation-pattern
    dtype: string
  - name: K-evaluatable
    dtype: bool
  - name: ground-truth
    dtype: string
  - name: sampled-statements
    list:
    - name: line_number
      dtype: int64
    - name: prior_state
      struct:
      - name: ble
        dtype: int64
      - name: ans
        dtype: int64
      - name: num
        dtype: int64
      - name: a
        dtype: int64
      - name: b
        dtype: int64
      - name: c
        dtype: int64
      - name: d
        dtype: int64
      - name: f
        dtype: int64
      - name: x
        dtype: int64
      - name: j
        dtype: int64
      - name: k
        dtype: int64
      - name: m
        dtype: int64
      - name: p
        dtype: int64
      - name: q
        dtype: int64
      - name: r
        dtype: int64
      - name: X
        dtype: int64
      - name: B
        dtype: int64
      - name: n
        dtype: int64
      - name: i
        dtype: int64
      - name: s
        dtype: int64
      - name: t
        dtype: int64
      - name: first
        dtype: int64
      - name: tk
        dtype: int64
      - name: y
        dtype: int64
      - name: T
        dtype: int64
      - name: aa
        dtype: int64
      - name: bb
        dtype: int64
      - name: tmp
        dtype: int64
      - name: xyz
        dtype: int64
      - name: e
        dtype: int64
      - name: l
        dtype: int64
      - name: h
        dtype: int64
      - name: count
        dtype: int64
      - name: x1
        dtype: int64
      - name: x2
        dtype: int64
      - name: y1
        dtype: int64
      - name: y2
        dtype: int64
      - name: sum
        dtype: int64
      - name: k9m
        dtype: int64
      - name: x7f
        dtype: int64
      - name: p2d
        dtype: int64
      - name: cnt
        dtype: int64
      - name: vol
        dtype: int64
      - name: z
        dtype: int64
      - name: n1
        dtype: int64
      - name: n2
        dtype: int64
      - name: l2
        dtype: int64
      - name: r2
        dtype: int64
      - name: temp
        dtype: int64
      - name: n3
        dtype: int64
      - name: number
        dtype: int64
      - name: ax1
        dtype: int64
      - name: ax2
        dtype: int64
      - name: ay1
        dtype: int64
      - name: ay2
        dtype: int64
      - name: bx1
        dtype: int64
      - name: bx2
        dtype: int64
      - name: by1
        dtype: int64
      - name: by2
        dtype: int64
      - name: left
        dtype: int64
      - name: right
        dtype: int64
      - name: oh
        dtype: int64
      - name: ow
        dtype: int64
      - name: r1x1
        dtype: int64
      - name: r1x2
        dtype: int64
      - name: r1y1
        dtype: int64
      - name: r1y2
        dtype: int64
      - name: r2x1
        dtype: int64
      - name: r2x2
        dtype: int64
      - name: r2y1
        dtype: int64
      - name: r2y2
        dtype: int64
    - name: rules
      list: string
    - name: cleaned_stmt
      dtype: string
    - name: control_stack
      dtype: string
  splits:
  - name: train
    num_bytes: 5013516
    num_examples: 324
  download_size: 187603
  dataset_size: 5013516
- config_name: predrule-IMP-K-uk-human-written
  features:
  - name: id
    dtype: string
  - name: program
    dtype: string
  - name: src-filename
    dtype: string
  - name: semantics-type
    dtype: string
  - name: syntax
    dtype: string
  - name: semantics
    dtype: string
  - name: language
    dtype: string
  - name: mutated-program
    dtype: string
  - name: mutated
    dtype: bool
  - name: mutation-pattern
    dtype: string
  - name: K-evaluatable
    dtype: bool
  - name: ground-truth
    dtype: string
  - name: sampled-statements
    list:
    - name: line_number
      dtype: int64
    - name: prior_state
      struct:
      - name: r1x1
        dtype: int64
      - name: ble
        dtype: int64
      - name: ans
        dtype: int64
      - name: r1x2
        dtype: int64
      - name: r1y1
        dtype: int64
      - name: r1y2
        dtype: int64
      - name: r2x1
        dtype: int64
      - name: r2x2
        dtype: int64
      - name: r2y1
        dtype: int64
      - name: r2y2
        dtype: int64
      - name: x
        dtype: int64
      - name: y
        dtype: int64
      - name: k
        dtype: int64
      - name: l
        dtype: int64
      - name: num
        dtype: int64
      - name: i
        dtype: int64
      - name: j
        dtype: int64
      - name: n
        dtype: int64
      - name: a
        dtype: int64
      - name: b
        dtype: int64
      - name: n1
        dtype: int64
      - name: n2
        dtype: int64
      - name: n3
        dtype: int64
      - name: c
        dtype: int64
      - name: d
        dtype: int64
      - name: f
        dtype: int64
      - name: m
        dtype: int64
      - name: p
        dtype: int64
      - name: q
        dtype: int64
      - name: r
        dtype: int64
      - name: X
        dtype: int64
      - name: B
        dtype: int64
      - name: s
        dtype: int64
      - name: t
        dtype: int64
      - name: first
        dtype: int64
      - name: tk
        dtype: int64
      - name: T
        dtype: int64
      - name: aa
        dtype: int64
      - name: bb
        dtype: int64
      - name: tmp
        dtype: int64
      - name: h
        dtype: int64
      - name: z
        dtype: int64
      - name: e
        dtype: int64
      - name: ax1
        dtype: int64
      - name: ax2
        dtype: int64
      - name: ay1
        dtype: int64
      - name: ay2
        dtype: int64
      - name: bx1
        dtype: int64
      - name: bx2
        dtype: int64
      - name: by1
        dtype: int64
      - name: by2
        dtype: int64
      - name: left
        dtype: int64
      - name: right
        dtype: int64
      - name: oh
        dtype: int64
      - name: ow
        dtype: int64
      - name: xyz
        dtype: int64
      - name: count
        dtype: int64
      - name: x1
        dtype: int64
      - name: x2
        dtype: int64
      - name: y1
        dtype: int64
      - name: y2
        dtype: int64
      - name: l2
        dtype: int64
      - name: r2
        dtype: int64
      - name: temp
        dtype: int64
      - name: number
        dtype: int64
      - name: sum
        dtype: int64
      - name: k9m
        dtype: int64
      - name: x7f
        dtype: int64
      - name: p2d
        dtype: int64
      - name: cnt
        dtype: int64
      - name: vol
        dtype: int64
    - name: rules
      list: string
    - name: cleaned_stmt
      dtype: string
    - name: control_stack
      dtype: string
  splits:
  - name: train
    num_bytes: 2478661
    num_examples: 162
  download_size: 129218
  dataset_size: 2478661
- config_name: predrule-IMP-SOS-mk-human-written
  features:
  - name: id
    dtype: string
  - name: program
    dtype: string
  - name: src-filename
    dtype: string
  - name: semantics-type
    dtype: string
  - name: syntax
    dtype: string
  - name: semantics
    dtype: string
  - name: language
    dtype: string
  - name: mutated-program
    dtype: string
  - name: mutated
    dtype: bool
  - name: mutation-pattern
    dtype: string
  - name: K-evaluatable
    dtype: bool
  - name: ground-truth
    dtype: string
  - name: sampled-statements
    list:
    - name: line_number
      dtype: int64
    - name: prior_state
      struct:
      - name: ble
        dtype: int64
      - name: ans
        dtype: int64
      - name: num
        dtype: int64
      - name: a
        dtype: int64
      - name: b
        dtype: int64
      - name: c
        dtype: int64
      - name: d
        dtype: int64
      - name: f
        dtype: int64
      - name: x
        dtype: int64
      - name: j
        dtype: int64
      - name: k
        dtype: int64
      - name: m
        dtype: int64
      - name: p
        dtype: int64
      - name: q
        dtype: int64
      - name: r
        dtype: int64
      - name: X
        dtype: int64
      - name: B
        dtype: int64
      - name: n
        dtype: int64
      - name: i
        dtype: int64
      - name: s
        dtype: int64
      - name: t
        dtype: int64
      - name: first
        dtype: int64
      - name: tk
        dtype: int64
      - name: y
        dtype: int64
      - name: T
        dtype: int64
      - name: aa
        dtype: int64
      - name: bb
        dtype: int64
      - name: tmp
        dtype: int64
      - name: xyz
        dtype: int64
      - name: e
        dtype: int64
      - name: l
        dtype: int64
      - name: h
        dtype: int64
      - name: x1
        dtype: int64
      - name: x2
        dtype: int64
      - name: y1
        dtype: int64
      - name: count
        dtype: int64
      - name: y2
        dtype: int64
      - name: sum
        dtype: int64
      - name: k9m
        dtype: int64
      - name: x7f
        dtype: int64
      - name: p2d
        dtype: int64
      - name: cnt
        dtype: int64
      - name: vol
        dtype: int64
      - name: z
        dtype: int64
      - name: temp
        dtype: int64
      - name: n1
        dtype: int64
      - name: n2
        dtype: int64
      - name: n3
        dtype: int64
      - name: l2
        dtype: int64
      - name: r2
        dtype: int64
      - name: number
        dtype: int64
      - name: ax1
        dtype: int64
      - name: ax2
        dtype: int64
      - name: ay1
        dtype: int64
      - name: ay2
        dtype: int64
      - name: bx1
        dtype: int64
      - name: bx2
        dtype: int64
      - name: by1
        dtype: int64
      - name: by2
        dtype: int64
      - name: left
        dtype: int64
      - name: oh
        dtype: int64
      - name: ow
        dtype: int64
      - name: right
        dtype: int64
      - name: r1x1
        dtype: int64
      - name: r1x2
        dtype: int64
      - name: r1y1
        dtype: int64
      - name: r1y2
        dtype: int64
      - name: r2x1
        dtype: int64
      - name: r2x2
        dtype: int64
      - name: r2y1
        dtype: int64
      - name: r2y2
        dtype: int64
    - name: rules
      list: string
    - name: cleaned_stmt
      dtype: string
    - name: control_stack
      dtype: string
  splits:
  - name: train
    num_bytes: 6783414
    num_examples: 324
  download_size: 208241
  dataset_size: 6783414
- config_name: predrule-IMP-SOS-uk-human-written
  features:
  - name: id
    dtype: string
  - name: program
    dtype: string
  - name: src-filename
    dtype: string
  - name: semantics-type
    dtype: string
  - name: syntax
    dtype: string
  - name: semantics
    dtype: string
  - name: language
    dtype: string
  - name: mutated-program
    dtype: string
  - name: mutated
    dtype: bool
  - name: mutation-pattern
    dtype: string
  - name: K-evaluatable
    dtype: bool
  - name: ground-truth
    dtype: string
  - name: sampled-statements
    list:
    - name: line_number
      dtype: int64
    - name: prior_state
      struct:
      - name: ans
        dtype: int64
      - name: ble
        dtype: int64
      - name: i
        dtype: int64
      - name: j
        dtype: int64
      - name: n
        dtype: int64
      - name: cnt
        dtype: int64
      - name: k
        dtype: int64
      - name: s
        dtype: int64
      - name: vol
        dtype: int64
      - name: p
        dtype: int64
      - name: x
        dtype: int64
      - name: a
        dtype: int64
      - name: b
        dtype: int64
      - name: y
        dtype: int64
      - name: z
        dtype: int64
      - name: c
        dtype: int64
      - name: d
        dtype: int64
      - name: m
        dtype: int64
      - name: r
        dtype: int64
      - name: q
        dtype: int64
      - name: xyz
        dtype: int64
      - name: e
        dtype: int64
      - name: h
        dtype: int64
      - name: l
        dtype: int64
      - name: temp
        dtype: int64
      - name: number
        dtype: int64
      - name: n1
        dtype: int64
      - name: n2
        dtype: int64
      - name: n3
        dtype: int64
      - name: f
        dtype: int64
      - name: x1
        dtype: int64
      - name: x2
        dtype: int64
      - name: y1
        dtype: int64
      - name: count
        dtype: int64
      - name: y2
        dtype: int64
      - name: sum
        dtype: int64
      - name: k9m
        dtype: int64
      - name: x7f
        dtype: int64
      - name: p2d
        dtype: int64
      - name: num
        dtype: int64
      - name: X
        dtype: int64
      - name: B
        dtype: int64
      - name: t
        dtype: int64
      - name: first
        dtype: int64
      - name: tk
        dtype: int64
      - name: T
        dtype: int64
      - name: l2
        dtype: int64
      - name: r2
        dtype: int64
      - name: aa
        dtype: int64
      - name: bb
        dtype: int64
      - name: tmp
        dtype: int64
      - name: ax1
        dtype: int64
      - name: ax2
        dtype: int64
      - name: ay1
        dtype: int64
      - name: ay2
        dtype: int64
      - name: bx1
        dtype: int64
      - name: bx2
        dtype: int64
      - name: by1
        dtype: int64
      - name: by2
        dtype: int64
      - name: left
        dtype: int64
      - name: oh
        dtype: int64
      - name: ow
        dtype: int64
      - name: right
        dtype: int64
      - name: r1x1
        dtype: int64
      - name: r1x2
        dtype: int64
      - name: r1y1
        dtype: int64
      - name: r1y2
        dtype: int64
      - name: r2x1
        dtype: int64
      - name: r2x2
        dtype: int64
      - name: r2y1
        dtype: int64
      - name: r2y2
        dtype: int64
    - name: rules
      list: string
    - name: cleaned_stmt
      dtype: string
    - name: control_stack
      dtype: string
  splits:
  - name: train
    num_bytes: 3362339
    num_examples: 162
  download_size: 142922
  dataset_size: 3362339
- config_name: predstate-IMP-K-mk-fuzzer-generated
  features:
  - name: id
    dtype: string
  - name: program
    dtype: string
  - name: src-filename
    dtype: string
  - name: semantics-type
    dtype: string
  - name: syntax
    dtype: string
  - name: semantics
    dtype: string
  - name: language
    dtype: string
  - name: mutated-program
    dtype: string
  - name: mutated
    dtype: bool
  - name: mutation-pattern
    dtype: string
  - name: K-evaluatable
    dtype: bool
  - name: ground-truth
    dtype: string
  splits:
  - name: train
    num_bytes: 26058090
    num_examples: 330
  download_size: 7001572
  dataset_size: 26058090
- config_name: predstate-IMP-K-mk-human-written
  features:
  - name: id
    dtype: string
  - name: program
    dtype: string
  - name: src-filename
    dtype: string
  - name: semantics-type
    dtype: string
  - name: syntax
    dtype: string
  - name: semantics
    dtype: string
  - name: language
    dtype: string
  - name: mutated-program
    dtype: string
  - name: mutated
    dtype: bool
  - name: mutation-pattern
    dtype: string
  - name: K-evaluatable
    dtype: bool
  - name: ground-truth
    dtype: string
  splits:
  - name: train
    num_bytes: 3235751
    num_examples: 324
  download_size: 63156
  dataset_size: 3235751
- config_name: predstate-IMP-K-mk-llm-translated
  features:
  - name: id
    dtype: string
  - name: program
    dtype: string
  - name: src-filename
    dtype: string
  - name: semantics-type
    dtype: string
  - name: syntax
    dtype: string
  - name: semantics
    dtype: string
  - name: language
    dtype: string
  - name: mutated-program
    dtype: string
  - name: mutated
    dtype: bool
  - name: mutation-pattern
    dtype: string
  - name: K-evaluatable
    dtype: bool
  - name: ground-truth
    dtype: string
  splits:
  - name: train
    num_bytes: 4354979
    num_examples: 330
  download_size: 268020
  dataset_size: 4354979
- config_name: predstate-IMP-K-uk-fuzzer-generated
  features:
  - name: id
    dtype: string
  - name: program
    dtype: string
  - name: src-filename
    dtype: string
  - name: semantics-type
    dtype: string
  - name: syntax
    dtype: string
  - name: semantics
    dtype: string
  - name: language
    dtype: string
  - name: mutated-program
    dtype: string
  - name: mutated
    dtype: bool
  - name: mutation-pattern
    dtype: string
  - name: K-evaluatable
    dtype: bool
  - name: ground-truth
    dtype: string
  splits:
  - name: train
    num_bytes: 12440461
    num_examples: 165
  download_size: 3420099
  dataset_size: 12440461
- config_name: predstate-IMP-K-uk-human-written
  features:
  - name: id
    dtype: string
  - name: program
    dtype: string
  - name: src-filename
    dtype: string
  - name: semantics-type
    dtype: string
  - name: syntax
    dtype: string
  - name: semantics
    dtype: string
  - name: language
    dtype: string
  - name: mutated-program
    dtype: string
  - name: mutated
    dtype: bool
  - name: mutation-pattern
    dtype: string
  - name: K-evaluatable
    dtype: bool
  - name: ground-truth
    dtype: string
  splits:
  - name: train
    num_bytes: 1591974
    num_examples: 162
  download_size: 45219
  dataset_size: 1591974
- config_name: predstate-IMP-K-uk-llm-translated
  features:
  - name: id
    dtype: string
  - name: program
    dtype: string
  - name: src-filename
    dtype: string
  - name: semantics-type
    dtype: string
  - name: syntax
    dtype: string
  - name: semantics
    dtype: string
  - name: language
    dtype: string
  - name: mutated-program
    dtype: string
  - name: mutated
    dtype: bool
  - name: mutation-pattern
    dtype: string
  - name: K-evaluatable
    dtype: bool
  - name: ground-truth
    dtype: string
  splits:
  - name: train
    num_bytes: 2135878
    num_examples: 165
  download_size: 191974
  dataset_size: 2135878
- config_name: predstate-IMP-SOS-mk-fuzzer-generated
  features:
  - name: id
    dtype: string
  - name: program
    dtype: string
  - name: src-filename
    dtype: string
  - name: semantics-type
    dtype: string
  - name: syntax
    dtype: string
  - name: semantics
    dtype: string
  - name: language
    dtype: string
  - name: mutated-program
    dtype: string
  - name: mutated
    dtype: bool
  - name: mutation-pattern
    dtype: string
  - name: K-evaluatable
    dtype: bool
  - name: ground-truth
    dtype: string
  splits:
  - name: train
    num_bytes: 27679877
    num_examples: 330
  download_size: 7008888
  dataset_size: 27679877
- config_name: predstate-IMP-SOS-mk-human-written
  features:
  - name: id
    dtype: string
  - name: program
    dtype: string
  - name: src-filename
    dtype: string
  - name: semantics-type
    dtype: string
  - name: syntax
    dtype: string
  - name: semantics
    dtype: string
  - name: language
    dtype: string
  - name: mutated-program
    dtype: string
  - name: mutated
    dtype: bool
  - name: mutation-pattern
    dtype: string
  - name: K-evaluatable
    dtype: bool
  - name: ground-truth
    dtype: string
  splits:
  - name: train
    num_bytes: 4828049
    num_examples: 324
  download_size: 70431
  dataset_size: 4828049
- config_name: predstate-IMP-SOS-mk-llm-translated
  features:
  - name: id
    dtype: string
  - name: program
    dtype: string
  - name: src-filename
    dtype: string
  - name: semantics-type
    dtype: string
  - name: syntax
    dtype: string
  - name: semantics
    dtype: string
  - name: language
    dtype: string
  - name: mutated-program
    dtype: string
  - name: mutated
    dtype: bool
  - name: mutation-pattern
    dtype: string
  - name: K-evaluatable
    dtype: bool
  - name: ground-truth
    dtype: string
  splits:
  - name: train
    num_bytes: 5976764
    num_examples: 330
  download_size: 275335
  dataset_size: 5976764
- config_name: predstate-IMP-SOS-uk-fuzzer-generated
  features:
  - name: id
    dtype: string
  - name: program
    dtype: string
  - name: src-filename
    dtype: string
  - name: semantics-type
    dtype: string
  - name: syntax
    dtype: string
  - name: semantics
    dtype: string
  - name: language
    dtype: string
  - name: mutated-program
    dtype: string
  - name: mutated
    dtype: bool
  - name: mutation-pattern
    dtype: string
  - name: K-evaluatable
    dtype: bool
  - name: ground-truth
    dtype: string
  splits:
  - name: train
    num_bytes: 13248959
    num_examples: 165
  download_size: 3426544
  dataset_size: 13248959
- config_name: predstate-IMP-SOS-uk-human-written
  features:
  - name: id
    dtype: string
  - name: program
    dtype: string
  - name: src-filename
    dtype: string
  - name: semantics-type
    dtype: string
  - name: syntax
    dtype: string
  - name: semantics
    dtype: string
  - name: language
    dtype: string
  - name: mutated-program
    dtype: string
  - name: mutated
    dtype: bool
  - name: mutation-pattern
    dtype: string
  - name: K-evaluatable
    dtype: bool
  - name: ground-truth
    dtype: string
  splits:
  - name: train
    num_bytes: 2385774
    num_examples: 162
  download_size: 51669
  dataset_size: 2385774
- config_name: predstate-IMP-SOS-uk-llm-translated
  features:
  - name: id
    dtype: string
  - name: program
    dtype: string
  - name: src-filename
    dtype: string
  - name: semantics-type
    dtype: string
  - name: syntax
    dtype: string
  - name: semantics
    dtype: string
  - name: language
    dtype: string
  - name: mutated-program
    dtype: string
  - name: mutated
    dtype: bool
  - name: mutation-pattern
    dtype: string
  - name: K-evaluatable
    dtype: bool
  - name: ground-truth
    dtype: string
  splits:
  - name: train
    num_bytes: 2944378
    num_examples: 165
  download_size: 198417
  dataset_size: 2944378
- config_name: predstate-IMP-nk-fuzzer-generated
  features:
  - name: id
    dtype: string
  - name: program
    dtype: string
  - name: src-filename
    dtype: string
  - name: language
    dtype: string
  - name: mutated-program
    dtype: string
  - name: mutated
    dtype: bool
  - name: mutation-pattern
    dtype: string
  - name: K-evaluatable
    dtype: bool
  - name: ground-truth
    dtype: string
  splits:
  - name: train
    num_bytes: 10917179
    num_examples: 165
  download_size: 3416057
  dataset_size: 10917179
- config_name: predstate-IMP-nk-human-written
  features:
  - name: id
    dtype: string
  - name: program
    dtype: string
  - name: src-filename
    dtype: string
  - name: language
    dtype: string
  - name: mutated-program
    dtype: string
  - name: mutated
    dtype: bool
  - name: mutation-pattern
    dtype: string
  - name: K-evaluatable
    dtype: bool
  - name: ground-truth
    dtype: string
  splits:
  - name: train
    num_bytes: 96390
    num_examples: 162
  download_size: 41198
  dataset_size: 96390
- config_name: predstate-IMP-nk-llm-translated
  features:
  - name: id
    dtype: string
  - name: program
    dtype: string
  - name: src-filename
    dtype: string
  - name: language
    dtype: string
  - name: mutated-program
    dtype: string
  - name: mutated
    dtype: bool
  - name: mutation-pattern
    dtype: string
  - name: K-evaluatable
    dtype: bool
  - name: ground-truth
    dtype: string
  splits:
  - name: train
    num_bytes: 612598
    num_examples: 165
  download_size: 187945
  dataset_size: 612598
- config_name: predtrace-IMP-K-mk-human-written
  features:
  - name: id
    dtype: string
  - name: program
    dtype: string
  - name: src-filename
    dtype: string
  - name: semantics-type
    dtype: string
  - name: syntax
    dtype: string
  - name: semantics
    dtype: string
  - name: language
    dtype: string
  - name: mutated-program
    dtype: string
  - name: mutated
    dtype: bool
  - name: mutation-pattern
    dtype: string
  - name: K-evaluatable
    dtype: bool
  - name: ground-truth
    dtype: string
  - name: max-loop-depth
    dtype: int64
  - name: max-if-depth
    dtype: int64
  splits:
  - name: train
    num_bytes: 22715525
    num_examples: 324
  download_size: 1272634
  dataset_size: 22715525
- config_name: predtrace-IMP-K-uk-human-written
  features:
  - name: id
    dtype: string
  - name: program
    dtype: string
  - name: src-filename
    dtype: string
  - name: semantics-type
    dtype: string
  - name: syntax
    dtype: string
  - name: semantics
    dtype: string
  - name: language
    dtype: string
  - name: mutated-program
    dtype: string
  - name: mutated
    dtype: bool
  - name: mutation-pattern
    dtype: string
  - name: K-evaluatable
    dtype: bool
  - name: ground-truth
    dtype: string
  - name: max-loop-depth
    dtype: int64
  - name: max-if-depth
    dtype: int64
  splits:
  - name: train
    num_bytes: 11331861
    num_examples: 162
  download_size: 648136
  dataset_size: 11331861
- config_name: predtrace-IMP-SOS-mk-human-written
  features:
  - name: id
    dtype: string
  - name: program
    dtype: string
  - name: src-filename
    dtype: string
  - name: semantics-type
    dtype: string
  - name: syntax
    dtype: string
  - name: semantics
    dtype: string
  - name: language
    dtype: string
  - name: mutated-program
    dtype: string
  - name: mutated
    dtype: bool
  - name: mutation-pattern
    dtype: string
  - name: K-evaluatable
    dtype: bool
  - name: ground-truth
    dtype: string
  - name: max-loop-depth
    dtype: int64
  - name: max-if-depth
    dtype: int64
  splits:
  - name: train
    num_bytes: 36623365
    num_examples: 324
  download_size: 2093312
  dataset_size: 36623365
- config_name: predtrace-IMP-SOS-uk-human-written
  features:
  - name: id
    dtype: string
  - name: program
    dtype: string
  - name: src-filename
    dtype: string
  - name: semantics-type
    dtype: string
  - name: syntax
    dtype: string
  - name: semantics
    dtype: string
  - name: language
    dtype: string
  - name: mutated-program
    dtype: string
  - name: mutated
    dtype: bool
  - name: mutation-pattern
    dtype: string
  - name: K-evaluatable
    dtype: bool
  - name: ground-truth
    dtype: string
  - name: max-loop-depth
    dtype: int64
  - name: max-if-depth
    dtype: int64
  splits:
  - name: train
    num_bytes: 18283431
    num_examples: 162
  download_size: 1061198
  dataset_size: 18283431
configs:
- config_name: predrule-IMP-K-mk-human-written
  data_files:
  - split: train
    path: predrule-IMP-K-mk-human-written/train-*
- config_name: predrule-IMP-K-uk-human-written
  data_files:
  - split: train
    path: predrule-IMP-K-uk-human-written/train-*
- config_name: predrule-IMP-SOS-mk-human-written
  data_files:
  - split: train
    path: predrule-IMP-SOS-mk-human-written/train-*
- config_name: predrule-IMP-SOS-uk-human-written
  data_files:
  - split: train
    path: predrule-IMP-SOS-uk-human-written/train-*
- config_name: predstate-IMP-K-mk-fuzzer-generated
  data_files:
  - split: train
    path: predstate-IMP-K-mk-fuzzer-generated/train-*
- config_name: predstate-IMP-K-mk-human-written
  data_files:
  - split: train
    path: predstate-IMP-K-mk-human-written/train-*
- config_name: predstate-IMP-K-mk-llm-translated
  data_files:
  - split: train
    path: predstate-IMP-K-mk-llm-translated/train-*
- config_name: predstate-IMP-K-uk-fuzzer-generated
  data_files:
  - split: train
    path: predstate-IMP-K-uk-fuzzer-generated/train-*
- config_name: predstate-IMP-K-uk-human-written
  data_files:
  - split: train
    path: predstate-IMP-K-uk-human-written/train-*
- config_name: predstate-IMP-K-uk-llm-translated
  data_files:
  - split: train
    path: predstate-IMP-K-uk-llm-translated/train-*
- config_name: predstate-IMP-SOS-mk-fuzzer-generated
  data_files:
  - split: train
    path: predstate-IMP-SOS-mk-fuzzer-generated/train-*
- config_name: predstate-IMP-SOS-mk-human-written
  data_files:
  - split: train
    path: predstate-IMP-SOS-mk-human-written/train-*
- config_name: predstate-IMP-SOS-mk-llm-translated
  data_files:
  - split: train
    path: predstate-IMP-SOS-mk-llm-translated/train-*
- config_name: predstate-IMP-SOS-uk-fuzzer-generated
  data_files:
  - split: train
    path: predstate-IMP-SOS-uk-fuzzer-generated/train-*
- config_name: predstate-IMP-SOS-uk-human-written
  data_files:
  - split: train
    path: predstate-IMP-SOS-uk-human-written/train-*
- config_name: predstate-IMP-SOS-uk-llm-translated
  data_files:
  - split: train
    path: predstate-IMP-SOS-uk-llm-translated/train-*
- config_name: predstate-IMP-nk-fuzzer-generated
  data_files:
  - split: train
    path: predstate-IMP-nk-fuzzer-generated/train-*
- config_name: predstate-IMP-nk-human-written
  data_files:
  - split: train
    path: predstate-IMP-nk-human-written/train-*
- config_name: predstate-IMP-nk-llm-translated
  data_files:
  - split: train
    path: predstate-IMP-nk-llm-translated/train-*
- config_name: predtrace-IMP-K-mk-human-written
  data_files:
  - split: train
    path: predtrace-IMP-K-mk-human-written/train-*
- config_name: predtrace-IMP-K-uk-human-written
  data_files:
  - split: train
    path: predtrace-IMP-K-uk-human-written/train-*
- config_name: predtrace-IMP-SOS-mk-human-written
  data_files:
  - split: train
    path: predtrace-IMP-SOS-mk-human-written/train-*
- config_name: predtrace-IMP-SOS-uk-human-written
  data_files:
  - split: train
    path: predtrace-IMP-SOS-uk-human-written/train-*
---

<div align="center">
  <p>The 43rd International Conference on Machine Learning (ICML 2026), Seoul, South Korea</p>
  <h1>
    LLMs Lean on Priors, Not Programming Language
    <span style="white-space: nowrap;">
      Semantics
      <img
        src="https://raw.githubusercontent.com/EngineeringSoftware/PLSemanticsBench/main/docs/logo.png"
        alt="PLSemanticsBench logo"
        width="60"
        style="display: inline-block !important;
               vertical-align: -1.0em;
               margin-left: 10px;
               margin-bottom: 0;">
    </span>
  </h1>

  <p style="font-size: 20px;">
    by
    <a href="https://www.adityathimmaiah.com">Aditya Thimmaiah</a><sup>1</sup>,
    <a href="https://jiyangzhang.github.io/">Jiyang Zhang</a><sup>1</sup>,
    <a href="https://scholar.google.com/citations?user=HtNfeKYAAAAJ&hl=en">Jayanth Srinivasa</a><sup>2</sup>,
    <a href="https://www.jessyli.com">Junyi Jessy Li</a><sup>1</sup>,
    <a href="https://users.ece.utexas.edu/~gligoric/">Milos Gligoric</a><sup>1</sup>
  </p>

  <p>
    <sup>1</sup>The University of Texas at Austin &nbsp;&nbsp;&nbsp;
    <sup>2</sup>Cisco Research
  </p>
</div>

<div align="center">

[![Website](https://img.shields.io/badge/Project_Page-PLSemanticsBench-blueviolet)](https://engineeringsoftware.github.io/PLSemanticsBench/)
[![arXiv](https://img.shields.io/badge/arXiv-2510.03415v3-b31b1b.svg)](https://arxiv.org/pdf/2510.03415v3)
[![Code](https://img.shields.io/badge/Code-GitHub-black)](https://github.com/EngineeringSoftware/PLSemanticsBench)
[![Dataset](https://img.shields.io/badge/🤗-Dataset-yellow)](https://huggingface.co/datasets/EngineeringSoftware/PLSemanticsBench)

</div>

---

**TLDR:** Frontier LLMs execute programs with up to 90–100% accuracy when symbols retain their usual meanings (e.g., + means addition). Under counterfactual semantic shifts (e.g., redefining + to mean subtraction) accuracy collapses by 40–70 percentage points. Despite handing the complete formal rules, the models keep answering as if the rules were never changed.
LLMs don't faithfully interpret the semantics they are given—they retrieve what symbols usually mean from pretraining.

## Abstract
Recent work asks whether large language models (LLMs) condition their reasoning on explicit rules rather than statistical regularities from pre-training. 
Program execution provides a canonical instance: formal semantics define behavior through symbolic transition rules that can be systematically altered 
under distribution shift. We investigate whether LLMs can condition their reasoning on formal semantics through program execution and introduce PLSEMANTICSBENCH, 
pairing featherweight C programs with two semantic systems—small-step operational semantics and K semantics—and probing four capabilities: composing rules for 
final states, selecting rules when state is unmutated, sustaining such conditioning over long traces, and following supplied rules under novel semantics. To 
decouple semantic reasoning from syntactic familiarity, we redefine familiar operators to induce symbol-meaning conflict and introduce novel symbols defined only
through the supplied rules, and stress-test models on Human-Written, LLM-Translated, and Fuzzer-Generated splits with increasing structural complexity. Across 
11 frontier LLMs, strong final-state accuracy under standard semantics (up to 90%) drops sharply—by as much as 40–60% points—under semantic mutations and increasing
structural complexity. Only a handful of models achieve non-zero long-horizon conditioning accuracy, and even the best systems reach just 35%. Together, these results 
suggest that contemporary LLMs often rely on pretrained lexical associations rather than systematically conditioning on supplied formal rules.

## Table of Contents
- [About](#about)
- [Installation](#installation)
- [Quick Start](#quick-start)
- [Detailed Usage](#detailed-usage)
- [Benchmark](#benchmark)
- [Citation](#citation)

## About
PLSemanticsBench is the first counterfactual PL semantics dataset for rule-conditioned reasoning of LLMs. We introduce three tasks to evaluate this:

| Task | Description |
|------|-------------|
| ✨ **PredState**| Predicts the final program state |
| ✨ **PredRule** | Predicts the ordered sequence of semantic rules needed to evaluate a program|
| ✨ **PredTrace**| Predicts the step-by-step execution of a program |

You must implement [BaseRunner](https://github.com/EngineeringSoftware/PLSemanticsBench/blob/main/src/plsemanticsbench/core/exps/base_experiment.py)(`_query` method) to evaluate your models. We provide two example implementations for OpenAI models ([GPTRunner](https://github.com/EngineeringSoftware/PLSemanticsBench/blob/main/src/plsemanticsbench/core/exps/gpt_experiment.py)) and Ollama models ([OllamaRunner](https://github.com/EngineeringSoftware/PLSemanticsBench/blob/main/src/plsemanticsbench/core/exps/ollama_experiment.py)).

## Installation

### System Requirements
- Python 3.11 or higher
- OpenAI API key (for running experiments with OpenAI models)


### Step-by-Step Installation
1. Create and activate the conda environment:
```bash
conda env create -f env.yaml
conda activate plsemanticsbench
```

2. Set up your OpenAI API key (only for OpenAI models):
```bash
export OPENAI_API_KEY='your-api-key-here'
```

## Quick Start

We provide a bash script `quick` that:
 1. Sets up the `plsemanticsbench` conda environment.
 2. Pulls the `DeepSeek-R1 1.5B` model.
 3. Evaluates the `DeepSeek-R1 1.5B` model on the `PredState` task with `no-semantics` and `chain-of-thought` prompting on the `Human-Written` dataset.
 4. Prints the `accuracy` and `malformed-count` to screen.
 5. Creates `metrics-predstate-deepseek-r1:1.5b.json` that contains the evaluation result.
 
```bash
bash quick
```

## Detailed Usage

### Basic Example
Here's a minimal example to get started:

```python
from plsemanticsbench import GPTRunner
from plsemanticsbench import ExperimentArgs, LLMEvaluator
from plsemanticsbench import (
    PROMPT_STRATEGY,
    Task,
    Formalization,
    Semantics_Type,
    Language,
    PLDataset
)

# Model name
model_name = "o3-mini"

# Experiment args: Run the PredState task on the IMP language with
# standard semantics formalized using SOS and with direct prompting
exp_args = ExperimentArgs(
    dataset=PLDataset.Human_Written,
    task=Task.PredState,
    language=Language.IMP,
    formalization=Formalization.SOS,
    semantics_type=Semantics_Type.Standard,
    model_name=model_name,
    prompt_strategy=PROMPT_STRATEGY.DA,
    num_datapoints_to_run=2, # Run just 2 datapoints (omit to run entire dataset)
)
                        
# Run inference using the OpenAI API
gpt_runner = GPTRunner(args=exp_args)

# Generation (generate LLM prediction on the predstate task)
predictions = gpt_runner.do_experiment() # path to dump results can be provided

# Evaluation (evaluate LLM prediction against ground-truth)
llm_eval = LLMEvaluator(task=exp_args.task, semantics_type=exp_args.semantics_type)
evaluation_result = llm_eval.evaluate_from_list(results=predictions, model_name=model_name)
print(evaluation_result)
```

### Expected Output

```python
{
    'accuracy': 1,
    'malformed-count': 0,
}
```

## Benchmark

You can load the dataset using the `datasets` library. Here is an example:
```python
from datasets import load_dataset

# Load PredState task with standard semantics (uk) and K-semantics formalization (K) and with the Human Written (human-written) dataset
predstate_IMP_K_uk_human_written = load_dataset("EngineeringSoftware/PLSemanticsBench", name="predstate-IMP-K-uk-human-written")

# Load PredRule task with nonstandard semantics (mk) ans SOS formalization (SOS) and with the LLM Translated (llm-translated) dataset
predrule_IMP_SOS_mk_llm_translated = load_dataset("EngineeringSoftware/PLSemanticsBench", name="predrule-IMP-SOS-mk-llm-translated")

# Load PredState task with no-semantics (nk) and with the Fuzzer Generated (fuzzer-generated) dataset
predstate_IMP_nk_fuzzer_generated = load_dataset("EngineeringSoftware/PLSemanticsBench", name="predstate-IMP-nk-fuzzer-generated")
```

### Dataset Split

<table>
  <tr>
    <th>Task</th>
    <th>Split</th>
    <th>Description</th>
  </tr>
  <tr>
    <td rowspan="5">✨ <strong>PredState</strong><br>(Final State Prediction)</td>
    <td> predstate-IMP-nk-{dataset-name} </td>
    <td> No semantics </td>
  </tr>
  <tr>
    <td> predstate-IMP-K-uk-{dataset-name} </td>
    <td>Standard semantics with K-semantics formalization</td>
  </tr>
  <tr>
    <td> predstate-IMP-K-mk-{dataset-name} </td>
    <td>Nonstandard semantics with K-semantics formalization</td>
  </tr>
  <tr>
    <td> predstate-IMP-SOS-uk-{dataset-name} </td>
    <td>Standard semantics with SOS formalization</td>
  </tr>
  <tr>
    <td> predstate-IMP-SOS-mk-{dataset-name} </td>
    <td>Nonstandard semantics with SOS formalization</td>
  </tr>
  <tr>
    <td rowspan="4">✨ <strong>PredRule</strong><br>(Semantic Rule Prediction)</td>
    <td> predrule-IMP-K-uk-human-written </td>
    <td>Standard semantics with K-semantics formalization</td>
  </tr>
  <tr>
    <td> predrule-IMP-K-mk-human-written </td>
    <td>Nonstandard semantics with K-semantics formalization</td>
  </tr>
  <tr>
    <td> predrule-IMP-SOS-uk-human-written </td>
    <td>Standard semantics with SOS formalization</td>
  </tr>
  <tr>
    <td> predrule-IMP-SOS-mk-human-written </td>
    <td>Nonstandard semantics with SOS formalization</td>
  </tr>
  <tr>
    <td rowspan="4">✨ <strong>PredTrace</strong><br>(Execution Trace Prediction)</td>
    <td> predtrace-IMP-K-uk-human-written </td>
    <td>Standard semantics with K-semantics formalization</td>
  </tr>
  <tr>
    <td> predtrace-IMP-K-mk-human-written </td>
    <td>Nonstandard semantics with K-semantics formalization</td>
  </tr>
  <tr>
    <td> predtrace-IMP-SOS-uk-human-written </td>
    <td>Standard semantics with SOS formalization</td>
  </tr>
  <tr>
    <td> predtrace-IMP-SOS-mk-human-written </td>
    <td>Nonstandard semantics with SOS formalization</td>
  </tr>
</table>


### Data Example

One example of the dataset is as follows:
```json
{
  "program": "int ans; ans = 1; ...",
  "syntax": "<program> :: ...",
  "semantics": "ℤ := Set of integers ...",
  "mutated-program": "int ans; ans = 1; ...",
  "mutation-pattern": "KeyWordSwap",
  "exec-trace": [
    {
      "linenumber": 1,
      "rule": ["Rule 38", "Rule 39"],
      "state": {"ans": 1}
    }
  ],
  "ground-truth": "<answer>...</answer>"
}
```

## Citation
```bibtex
@inproceedings{ThimmaiahETAL25PLSemanticsBench,
  title     = {LLMs Lean on Priors, Not Programming Language Semantics},
  author    = {Thimmaiah, Aditya and Zhang, Jiyang and Srinivasa, Jayanth
               and Li, Junyi Jessy and Gligoric, Milos},
  booktitle = {ICML},
  year      = {2026}
}
```


## License
This project is licensed under the [CC BY 4.0 License](https://creativecommons.org/licenses/by/4.0/).