Update README.md
Browse files
README.md
CHANGED
@@ -324,7 +324,7 @@ configs:
|
|
324 |
# Easy2Hard-Bench
|
325 |
|
326 |
<div align="center">
|
327 |
-
<img src="./img/hf_data.png" alt="Logo" width="
|
328 |
</div>
|
329 |
|
330 |
## Dataset Description
|
@@ -404,8 +404,8 @@ DatasetDict({
|
|
404 |
})
|
405 |
```
|
406 |
|
407 |
-
|
408 |
-
|
409 |
|Field|Type|Description|
|
410 |
|---|---|---|
|
411 |
|contest|string|name of the contest|
|
@@ -429,7 +429,7 @@ DatasetDict({
|
|
429 |
|unnorm_rating_upper|float|upper threshold of difficulty suggested by AoPS|
|
430 |
|ever_exist|bool|whether the problem exists in the MATH dataset|
|
431 |
|
432 |
-
|
433 |
|Field|Type|Description|
|
434 |
|---|---|---|
|
435 |
|contest_id|int|Codeforce contest id|
|
@@ -467,9 +467,7 @@ DatasetDict({
|
|
467 |
|original_tags|string|original tags on Codeforces|
|
468 |
|ever_exist|bool|whether the problem exists in the APPS dataset|
|
469 |
|
470 |
-
|
471 |
-
|
472 |
-
#### E2H-Lichess
|
473 |
|Field|Type|Description|
|
474 |
|---|---|---|
|
475 |
|puzzle_id|string|id of the puzzle on Lichess|
|
@@ -503,7 +501,7 @@ If the number of solutions is less than 3, the data fields related to Solution 1
|
|
503 |
|previous_fen|string|Forsyth–Edwards notation (FEN) of the puzzle before last move by the opponent|
|
504 |
|last_move_uci|string|universal chess interface (UCI) notation of last move by the opponent|
|
505 |
|
506 |
-
|
507 |
Besides the data fields from the original datasets, all of these three datasets have the following difficulty-realted data fields:
|
508 |
|
509 |
|Field|Type|Description|
|
@@ -515,7 +513,7 @@ Besides the data fields from the original datasets, all of these three datasets
|
|
515 |
|unnorm_rating|float|unnormalized estimated difficulty|
|
516 |
|unnorm_rating_std|float|standard deviation of unnormalized estimated difficulty|
|
517 |
|
518 |
-
|
519 |
|
520 |
For the newly crafted datasets, E2H-AMC, E2H-Codeforces and E2H-Lichess, all of them contain a train and evaluation splits.
|
521 |
|
@@ -530,10 +528,10 @@ For the datasets, E2H-GSM8K, E2H-ARC and E2H-Winogrande, all of them only have e
|
|
530 |
| E2H-ARC | N.A. | 1,172 |
|
531 |
| E2H-Winogrande | N.A. | 1,267 |
|
532 |
|
533 |
-
|
534 |
|
535 |
<div align="center">
|
536 |
-
<img src="./img/hf_distribution.png" alt="Logo" width="
|
537 |
</div>
|
538 |
|
539 |
## Dataset Creation
|
@@ -548,4 +546,4 @@ For the datasets, E2H-GSM8K, E2H-ARC and E2H-Winogrande, all of them only have e
|
|
548 |
|
549 |
```
|
550 |
TBD
|
551 |
-
```
|
|
|
324 |
# Easy2Hard-Bench
|
325 |
|
326 |
<div align="center">
|
327 |
+
<img src="./img/hf_data.png" alt="Logo" width="95%">
|
328 |
</div>
|
329 |
|
330 |
## Dataset Description
|
|
|
404 |
})
|
405 |
```
|
406 |
|
407 |
+
## Data Fields
|
408 |
+
### E2H-AMC
|
409 |
|Field|Type|Description|
|
410 |
|---|---|---|
|
411 |
|contest|string|name of the contest|
|
|
|
429 |
|unnorm_rating_upper|float|upper threshold of difficulty suggested by AoPS|
|
430 |
|ever_exist|bool|whether the problem exists in the MATH dataset|
|
431 |
|
432 |
+
### E2H-Codeforces
|
433 |
|Field|Type|Description|
|
434 |
|---|---|---|
|
435 |
|contest_id|int|Codeforce contest id|
|
|
|
467 |
|original_tags|string|original tags on Codeforces|
|
468 |
|ever_exist|bool|whether the problem exists in the APPS dataset|
|
469 |
|
470 |
+
### E2H-Lichess
|
|
|
|
|
471 |
|Field|Type|Description|
|
472 |
|---|---|---|
|
473 |
|puzzle_id|string|id of the puzzle on Lichess|
|
|
|
501 |
|previous_fen|string|Forsyth–Edwards notation (FEN) of the puzzle before last move by the opponent|
|
502 |
|last_move_uci|string|universal chess interface (UCI) notation of last move by the opponent|
|
503 |
|
504 |
+
### E2H-GSM8K, E2H-ARC, E2H-Winogrande
|
505 |
Besides the data fields from the original datasets, all of these three datasets have the following difficulty-realted data fields:
|
506 |
|
507 |
|Field|Type|Description|
|
|
|
513 |
|unnorm_rating|float|unnormalized estimated difficulty|
|
514 |
|unnorm_rating_std|float|standard deviation of unnormalized estimated difficulty|
|
515 |
|
516 |
+
## Data Splits
|
517 |
|
518 |
For the newly crafted datasets, E2H-AMC, E2H-Codeforces and E2H-Lichess, all of them contain a train and evaluation splits.
|
519 |
|
|
|
528 |
| E2H-ARC | N.A. | 1,172 |
|
529 |
| E2H-Winogrande | N.A. | 1,267 |
|
530 |
|
531 |
+
## Data Difficulty Distribution
|
532 |
|
533 |
<div align="center">
|
534 |
+
<img src="./img/hf_distribution.png" alt="Logo" width="75%">
|
535 |
</div>
|
536 |
|
537 |
## Dataset Creation
|
|
|
546 |
|
547 |
```
|
548 |
TBD
|
549 |
+
```
|