Dataset Viewer
Auto-converted to Parquet Duplicate
board
listlengths
9
9
scramble_depth
int64
1
31
bucket
stringclasses
3 values
action_interface
stringclasses
1 value
optimal_actions
listlengths
1
31
optimal_length
int64
1
31
[ 1, 2, 3, 4, 5, 0, 7, 8, 6 ]
1
easy
tile_id_v1
[ 6 ]
1
[ 1, 2, 3, 4, 5, 6, 7, 0, 8 ]
1
easy
tile_id_v1
[ 8 ]
1
[ 1, 2, 0, 4, 5, 3, 7, 8, 6 ]
2
easy
tile_id_v1
[ 3, 6 ]
2
[ 1, 2, 3, 4, 0, 6, 7, 5, 8 ]
2
easy
tile_id_v1
[ 5, 8 ]
2
[ 1, 0, 3, 4, 2, 5, 7, 8, 6 ]
3
easy
tile_id_v1
[ 2, 5, 6 ]
3
[ 1, 2, 3, 0, 4, 5, 7, 8, 6 ]
3
easy
tile_id_v1
[ 4, 5, 6 ]
3
[ 0, 2, 3, 1, 4, 5, 7, 8, 6 ]
4
easy
tile_id_v1
[ 1, 4, 5, 6 ]
4
[ 1, 2, 3, 5, 0, 6, 4, 7, 8 ]
4
easy
tile_id_v1
[ 5, 4, 7, 8 ]
4
[ 1, 2, 3, 0, 8, 5, 4, 7, 6 ]
5
easy
tile_id_v1
[ 4, 7, 8, 5, 6 ]
5
[ 4, 1, 2, 0, 5, 3, 7, 8, 6 ]
5
easy
tile_id_v1
[ 4, 1, 2, 3, 6 ]
5
[ 0, 5, 2, 1, 4, 3, 7, 8, 6 ]
6
easy
tile_id_v1
[ 1, 4, 5, 2, 3, 6 ]
6
[ 2, 3, 0, 1, 5, 6, 4, 7, 8 ]
6
easy
tile_id_v1
[ 3, 2, 1, 4, 7, 8 ]
6
[ 1, 0, 2, 5, 6, 3, 4, 7, 8 ]
7
easy
tile_id_v1
[ 2, 3, 6, 5, 4, 7, 8 ]
7
[ 1, 0, 3, 8, 2, 5, 4, 7, 6 ]
7
easy
tile_id_v1
[ 2, 8, 4, 7, 8, 5, 6 ]
7
[ 1, 2, 3, 5, 6, 8, 0, 4, 7 ]
8
easy
tile_id_v1
[ 4, 7, 8, 6, 5, 4, 7, 8 ]
8
[ 1, 2, 3, 7, 0, 4, 5, 8, 6 ]
8
easy
tile_id_v1
[ 4, 6, 8, 5, 7, 4, 5, 8 ]
8
[ 1, 0, 2, 7, 6, 3, 5, 4, 8 ]
9
easy
tile_id_v1
[ 2, 3, 6, 4, 5, 7, 4, 5, 8 ]
9
[ 4, 1, 3, 7, 5, 0, 8, 2, 6 ]
9
easy
tile_id_v1
[ 5, 2, 8, 7, 4, 1, 2, 5, 6 ]
9
[ 1, 3, 6, 5, 7, 2, 0, 4, 8 ]
10
easy
tile_id_v1
[ 4, 7, 2, 6, 3, 2, 5, 4, 7, 8 ]
10
[ 4, 1, 3, 2, 0, 8, 7, 6, 5 ]
10
easy
tile_id_v1
[ 8, 5, 6, 8, 2, 4, 1, 2, 5, 6 ]
10
[ 1, 6, 2, 4, 8, 5, 7, 0, 3 ]
11
medium
tile_id_v1
[ 8, 5, 3, 8, 5, 6, 2, 3, 6, 5, 8 ]
11
[ 4, 0, 1, 7, 2, 3, 5, 8, 6 ]
11
medium
tile_id_v1
[ 1, 3, 6, 8, 5, 7, 4, 1, 2, 5, 8 ]
11
[ 1, 3, 6, 7, 0, 8, 2, 4, 5 ]
12
medium
tile_id_v1
[ 4, 2, 7, 4, 2, 5, 8, 6, 3, 2, 5, 8 ]
12
[ 3, 4, 5, 7, 1, 2, 0, 8, 6 ]
12
medium
tile_id_v1
[ 7, 1, 4, 3, 1, 4, 2, 5, 3, 2, 5, 6 ]
12
[ 1, 0, 2, 7, 4, 3, 6, 5, 8 ]
13
medium
tile_id_v1
[ 2, 3, 8, 5, 6, 7, 4, 8, 5, 6, 8, 5, 6 ]
13
[ 1, 6, 2, 3, 8, 0, 4, 7, 5 ]
13
medium
tile_id_v1
[ 8, 3, 4, 7, 5, 8, 3, 6, 2, 3, 6, 5, 8 ]
13
[ 2, 4, 1, 7, 0, 3, 8, 5, 6 ]
14
medium
tile_id_v1
[ 5, 8, 7, 2, 4, 1, 3, 5, 2, 4, 1, 2, 5, 6 ]
14
[ 3, 4, 6, 1, 8, 5, 7, 2, 0 ]
14
medium
tile_id_v1
[ 5, 8, 4, 3, 1, 4, 2, 5, 8, 6, 3, 2, 5, 8 ]
14
[ 1, 3, 6, 7, 5, 2, 8, 0, 4 ]
15
medium
tile_id_v1
[ 5, 2, 6, 3, 2, 5, 4, 6, 5, 4, 8, 7, 4, 5, 6 ]
15
[ 1, 8, 6, 4, 3, 2, 7, 0, 5 ]
15
medium
tile_id_v1
[ 3, 8, 6, 2, 8, 3, 5, 8, 3, 6, 2, 3, 6, 5, 8 ]
15
[ 1, 2, 5, 6, 0, 7, 4, 8, 3 ]
16
medium
tile_id_v1
[ 7, 3, 8, 7, 6, 1, 2, 5, 3, 6, 5, 2, 1, 4, 7, 8 ]
16
[ 8, 5, 1, 2, 7, 3, 4, 6, 0 ]
16
medium
tile_id_v1
[ 6, 7, 5, 1, 3, 5, 2, 8, 1, 2, 8, 4, 7, 8, 5, 6 ]
16
[ 1, 5, 2, 8, 6, 0, 3, 7, 4 ]
17
medium
tile_id_v1
[ 6, 7, 3, 8, 7, 3, 4, 6, 3, 4, 8, 7, 4, 5, 2, 3, 6 ]
17
[ 3, 7, 6, 2, 4, 0, 5, 1, 8 ]
17
medium
tile_id_v1
[ 4, 7, 3, 2, 7, 1, 5, 7, 1, 4, 6, 3, 2, 1, 4, 5, 8 ]
17
[ 0, 6, 8, 3, 2, 1, 7, 4, 5 ]
18
medium
tile_id_v1
[ 3, 2, 1, 8, 6, 3, 2, 1, 4, 5, 8, 6, 3, 2, 1, 4, 5, 8 ]
18
[ 4, 3, 7, 1, 0, 2, 5, 8, 6 ]
18
medium
tile_id_v1
[ 3, 7, 2, 3, 7, 4, 1, 7, 4, 2, 3, 6, 8, 5, 7, 4, 5, 8 ]
18
[ 1, 6, 8, 4, 7, 0, 2, 3, 5 ]
19
medium
tile_id_v1
[ 8, 6, 1, 4, 7, 3, 2, 7, 4, 1, 3, 2, 5, 8, 6, 3, 2, 5, 8 ]
19
[ 4, 3, 6, 0, 5, 1, 7, 8, 2 ]
19
medium
tile_id_v1
[ 7, 8, 5, 1, 6, 3, 1, 5, 2, 6, 5, 2, 8, 7, 4, 1, 2, 5, 6 ]
19
[ 3, 6, 4, 7, 2, 5, 0, 8, 1 ]
20
medium
tile_id_v1
[ 7, 2, 5, 1, 8, 5, 1, 4, 6, 3, 2, 1, 4, 6, 3, 2, 1, 4, 5, 8 ]
20
[ 5, 1, 0, 4, 6, 8, 7, 3, 2 ]
20
medium
tile_id_v1
[ 1, 6, 3, 2, 8, 3, 6, 1, 3, 6, 2, 7, 4, 5, 1, 2, 5, 4, 7, 8 ]
20
[ 1, 5, 7, 0, 2, 6, 4, 3, 8 ]
21
hard
tile_id_v1
[ 2, 3, 8, 6, 7, 5, 3, 7, 6, 8, 7, 2, 4, 7, 8, 6, 5, 3, 2, 5, 6 ]
21
[ 7, 5, 6, 2, 3, 4, 8, 0, 1 ]
21
hard
tile_id_v1
[ 1, 4, 6, 5, 3, 1, 4, 6, 5, 3, 1, 2, 7, 1, 2, 4, 8, 7, 4, 5, 6 ]
21
[ 0, 7, 6, 5, 2, 3, 1, 8, 4 ]
22
hard
tile_id_v1
[ 7, 2, 3, 4, 8, 1, 5, 7, 2, 3, 1, 5, 7, 1, 4, 6, 3, 2, 1, 4, 5, 8 ]
22
[ 6, 2, 8, 5, 7, 1, 4, 3, 0 ]
22
hard
tile_id_v1
[ 3, 7, 1, 8, 2, 6, 5, 1, 6, 5, 1, 4, 7, 6, 8, 3, 6, 8, 5, 2, 3, 6 ]
22
[ 2, 1, 5, 7, 8, 0, 6, 4, 3 ]
23
hard
tile_id_v1
[ 8, 4, 6, 7, 4, 8, 3, 6, 8, 1, 5, 3, 6, 8, 7, 4, 1, 5, 2, 1, 4, 7, 8 ]
23
[ 6, 0, 4, 8, 7, 3, 2, 1, 5 ]
23
hard
tile_id_v1
[ 6, 8, 2, 1, 7, 6, 8, 2, 1, 7, 6, 8, 4, 3, 5, 6, 8, 4, 2, 1, 4, 5, 6 ]
23
[ 3, 5, 0, 6, 2, 1, 7, 4, 8 ]
24
hard
tile_id_v1
[ 5, 3, 6, 2, 1, 5, 3, 6, 2, 1, 6, 3, 5, 6, 4, 8, 6, 5, 3, 2, 1, 4, 5, 6 ]
24
[ 6, 8, 0, 3, 5, 2, 4, 1, 7 ]
24
hard
tile_id_v1
[ 2, 5, 1, 7, 5, 2, 8, 6, 3, 1, 2, 8, 6, 3, 1, 4, 7, 5, 8, 6, 3, 2, 5, 8 ]
24
[ 6, 7, 4, 3, 2, 0, 8, 1, 5 ]
25
hard
tile_id_v1
[ 2, 3, 6, 7, 4, 2, 3, 1, 8, 6, 7, 4, 1, 8, 6, 7, 4, 1, 2, 3, 5, 6, 8, 5, 6 ]
25
[ 8, 1, 5, 6, 2, 4, 3, 0, 7 ]
25
hard
tile_id_v1
[ 7, 4, 2, 6, 8, 1, 5, 2, 6, 7, 3, 8, 7, 3, 4, 6, 3, 4, 8, 7, 4, 5, 2, 3, 6 ]
25
[ 6, 7, 0, 1, 3, 5, 2, 8, 4 ]
26
hard
tile_id_v1
[ 5, 3, 7, 6, 1, 7, 3, 5, 6, 3, 8, 4, 5, 8, 4, 2, 7, 4, 2, 5, 8, 6, 3, 2, 5, 8 ]
26
[ 8, 7, 6, 2, 3, 4, 0, 5, 1 ]
26
hard
tile_id_v1
[ 2, 8, 7, 3, 4, 6, 3, 4, 5, 1, 6, 5, 1, 2, 8, 7, 4, 1, 2, 8, 7, 4, 1, 2, 5, 6 ]
26
[ 6, 0, 5, 2, 3, 7, 8, 4, 1 ]
27
hard
tile_id_v1
[ 3, 7, 5, 3, 6, 2, 7, 6, 3, 5, 6, 4, 1, 6, 4, 1, 8, 7, 1, 4, 5, 3, 2, 1, 4, 5, 6 ]
27
[ 7, 8, 5, 1, 4, 6, 2, 0, 3 ]
27
hard
tile_id_v1
[ 4, 8, 7, 1, 2, 4, 8, 7, 5, 6, 3, 8, 7, 3, 6, 5, 3, 2, 4, 7, 8, 6, 5, 3, 2, 5, 6 ]
27
[ 0, 8, 6, 7, 4, 3, 1, 2, 5 ]
28
hard
tile_id_v1
[ 8, 6, 3, 5, 2, 4, 7, 1, 4, 7, 6, 8, 1, 4, 7, 6, 8, 3, 5, 2, 6, 8, 2, 5, 3, 2, 5, 6 ]
28
[ 7, 5, 0, 4, 8, 6, 2, 3, 1 ]
28
hard
tile_id_v1
[ 5, 7, 4, 8, 6, 1, 3, 6, 7, 5, 1, 3, 6, 7, 8, 2, 7, 8, 5, 1, 3, 5, 2, 4, 1, 2, 5, 6 ]
28
[ 2, 4, 7, 0, 5, 8, 3, 6, 1 ]
29
hard
tile_id_v1
[ 5, 6, 3, 5, 2, 4, 6, 3, 1, 8, 7, 6, 3, 7, 6, 3, 4, 2, 7, 1, 5, 7, 1, 4, 2, 1, 4, 5, 8 ]
29
[ 2, 8, 7, 3, 5, 0, 6, 4, 1 ]
29
hard
tile_id_v1
[ 7, 8, 5, 3, 6, 4, 1, 7, 8, 5, 3, 6, 4, 1, 7, 8, 6, 4, 1, 7, 8, 6, 5, 3, 2, 1, 4, 5, 6 ]
29
[ 0, 7, 8, 2, 5, 4, 3, 6, 1 ]
30
hard
tile_id_v1
[ 2, 5, 7, 8, 4, 1, 6, 7, 8, 2, 5, 3, 7, 8, 2, 4, 1, 2, 3, 5, 4, 1, 2, 3, 5, 4, 1, 2, 3, 6 ]
30
[ 6, 5, 7, 8, 4, 2, 3, 1, 0 ]
30
hard
tile_id_v1
[ 2, 7, 5, 4, 1, 3, 8, 6, 4, 1, 7, 2, 3, 7, 6, 8, 7, 6, 8, 4, 1, 5, 2, 3, 6, 8, 5, 2, 3, 6 ]
30
[ 6, 4, 7, 8, 5, 0, 3, 2, 1 ]
31
hard
tile_id_v1
[ 1, 2, 3, 8, 6, 4, 7, 1, 2, 3, 5, 6, 8, 5, 6, 7, 1, 2, 3, 6, 7, 8, 5, 7, 8, 5, 4, 1, 2, 3, 6 ]
31
[ 8, 6, 7, 2, 5, 4, 3, 0, 1 ]
31
hard
tile_id_v1
[ 5, 6, 7, 4, 6, 2, 3, 5, 1, 6, 2, 3, 8, 7, 4, 2, 3, 1, 5, 8, 7, 4, 1, 5, 8, 7, 4, 1, 2, 3, 6 ]
31

8-puzzle / 3x3 sliding puzzle

Fixed 3x3 / 8-puzzle evaluation data and teacher-rollout datasets for sliding-puzzle reasoning experiments.

Configs

  • eval: 62 fixed evaluation puzzles, with exactly two puzzles at every optimal distance from 1 through 31.
  • sft-source: 200 fresh boards, balanced with 20 boards at each exact depth 10-19. It excludes the reserved evaluation boards and deliberately contains no solution-action labels.
  • sft: 3,745 replay-verified, fully annotated teacher decisions from solved sft-source trajectories.

The eval configuration uses the eval split. sft-source and sft use the train split.

Fields

  • board: flat 3x3 board, with 0 as the blank tile.
  • scramble_depth: exact optimal solution length for the generated evaluation set.
  • bucket: easy, medium, or hard.
  • action_interface: schema identifier, tile_id_v1.
  • optimal_actions: integer tile IDs to slide from the board to the solved state.
  • optimal_length: length of optimal_actions.

Evaluation set

The evaluation set contains two puzzles at every exact optimal distance from 1 through 31. The boards are sampled reproducibly from the reachable state space and are kept separate from the local SFT datasets. Every stored action sequence is replay-validated and optimal.

Bucket Optimal length Count
easy 1-10 20
medium 11-20 20
hard 21-31 22

Actions are numbered tiles from 1 through 8; each action slides that tile into the blank.

sft-source fields

  • id: stable source-board identifier.
  • board: flat 3x3 board, with 0 as the blank tile.
  • optimal_length: exact optimal solution length; source metadata only and not a teacher-model input.
  • action_interface: schema identifier, tile_id_v1.

sft-source intentionally omits optimal_actions; teacher models must produce their own trajectories.

sft fields

  • prompt: evaluator-format chat messages containing the system prompt, current board, and up to four prior rationale/action/result turns.
  • completion: one assistant message containing the annotated rationale and one verified slide_tile tool call.
  • tools: the slide_tile function schema used by the evaluator.
  • metadata: replay and audit fields including source_id, target_turn, board, tile, next_board, and history_turns.

The SFT training records were generated with a fixed history window of four turns.

Downloads last month
443