task string | model string | model_key string | data_id int32 | seed int32 | prompt string | trigger string | sampling dict | text string | finish_reason string | num_prompt_tokens int32 | num_completion_tokens int32 | ground_truth string | schema_version int32 | evalscope_extracted_answer string | evalscope_is_correct int8 | grading_status string | grader_id string | extracted_answer string | extracted_answer_is_list bool | boxed_answer_raw string | has_box int8 | has_valid_box int8 | num_boxes int32 | box_status string | boxed_is_correct int8 | evalscope_extraction_method string | cv3b_label string | cv3b_label_meaning string | cv3b_prob float64 | cv3b_abc_A float64 | cv3b_abc_B float64 | cv3b_abc_C float64 | cv3b_ctx_A float64 | cv3b_ctx_B float64 | cv3b_ctx_C float64 | llmv_model string | llmv_model_revision string | llmv_problem_understanding_expected float64 | llmv_reasoning_validity_expected float64 | llmv_conclusion_support_expected float64 | llmv_mean_score_token_entropy_nats float64 | llmv_mean_scale_probability_mass_before_renormalization float64 | llmv_n_fine_grained_scores int32 | llmv_score_definition string | logprob_sentinel bool | tokens dict |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
aime_2026 | Qwen/Qwen3.6-35B-A3B | Qwen3.6-35B-A3B | 0 | 0 | Patrick started walking at a constant rate along a straight road from school to the park. One hour after Patrick left, Tanya started running along the same road from school to the park. One hour after Tanya left, Jose started bicycling along the same road from school to the park. Tanya ran at a constant rate of $2$ mil... |
Solve the following math problem step by step. Put your answer inside \boxed{}.
{question}
Remember to put your answer inside \boxed{}. | {
"strategy": "top_p",
"temperature": 1,
"top_p": 0.95,
"top_k": 20,
"min_p": 0,
"presence_penalty": 1.5,
"repetition_penalty": 1,
"n": 1,
"max_tokens": 81920,
"skip_special_tokens": false,
"reasoning_effort": null,
"seed": 0
} | Here's a thinking process to solve the problem:
1. **Understand the Goal:** The objective is to find the distance from school to the park, expressed as a fraction $m/n$, and then calculate $m+n$. The problem involves three people (Patrick, Tanya, Jose) traveling at different speeds and starting at different times, bu... | stop | 186 | 7,206 | 277 | 1 | 277 | 1 | ok | evalscope-1.8.1-aime-rule-v1 | 277 | false | 277 | 1 | 1 | 1 | ok | 1 | boxed | A | correct | 0.999934 | 0.999947 | 0.000051 | 0.000002 | 0.404471 | 0.404471 | 0.191058 | Qwen/Qwen3.6-35B-A3B | 995ad96eacd98c81ed38be0c5b274b04031597b0 | 16.754517 | 19.705077 | 19.124362 | 0.279957 | 0.995849 | 3 | mean over three criterion-specific pointwise expected A--T score-token distributions; normalized to [0,1] | false | {
"prompt_sum_logprob": -456.3643849330583,
"prompt_avg_logprob": -2.4668345131516665,
"prompt_ppl": 11.785082221445515,
"completion_sum_logprob": -1253.5036824007116,
"completion_avg_logprob": -0.17395277302258,
"completion_ppl": 1.1899993644201488
} |
aime_2026 | Qwen/Qwen3.6-35B-A3B | Qwen3.6-35B-A3B | 0 | 1 | Patrick started walking at a constant rate along a straight road from school to the park. One hour after Patrick left, Tanya started running along the same road from school to the park. One hour after Tanya left, Jose started bicycling along the same road from school to the park. Tanya ran at a constant rate of $2$ mil... |
Solve the following math problem step by step. Put your answer inside \boxed{}.
{question}
Remember to put your answer inside \boxed{}. | {
"strategy": "top_p",
"temperature": 1,
"top_p": 0.95,
"top_k": 20,
"min_p": 0,
"presence_penalty": 1.5,
"repetition_penalty": 1,
"n": 1,
"max_tokens": 81920,
"skip_special_tokens": false,
"reasoning_effort": null,
"seed": 1
} | Here's a thinking process to solve the problem:
1. **Understand the Goal:** The objective is to find the distance from school to the park, expressed as a fraction $\frac{m}{n}$ in simplest form, and then calculate $m+n$.
2. **Identify the Variables and Relationships:**
* **People:** Patrick (P), Tanya (T), Jo... | stop | 186 | 8,766 | 277 | 1 | 277 | 1 | ok | evalscope-1.8.1-aime-rule-v1 | 277 | false | 277 | 1 | 1 | 1 | ok | 1 | boxed | A | correct | 0.999895 | 0.999902 | 0.000096 | 0.000002 | 0.489621 | 0.381317 | 0.129063 | Qwen/Qwen3.6-35B-A3B | 995ad96eacd98c81ed38be0c5b274b04031597b0 | 19.114918 | 19.271282 | 18.965031 | 0.225135 | 0.998018 | 3 | mean over three criterion-specific pointwise expected A--T score-token distributions; normalized to [0,1] | false | {
"prompt_sum_logprob": -470.98194338182657,
"prompt_avg_logprob": -2.545848342604468,
"prompt_ppl": 12.75404329622253,
"completion_sum_logprob": -1473.8150084849622,
"completion_avg_logprob": -0.1681285658778191,
"completion_ppl": 1.1830887057084711
} |
aime_2026 | Qwen/Qwen3.6-35B-A3B | Qwen3.6-35B-A3B | 0 | 2 | Patrick started walking at a constant rate along a straight road from school to the park. One hour after Patrick left, Tanya started running along the same road from school to the park. One hour after Tanya left, Jose started bicycling along the same road from school to the park. Tanya ran at a constant rate of $2$ mil... |
Solve the following math problem step by step. Put your answer inside \boxed{}.
{question}
Remember to put your answer inside \boxed{}. | {
"strategy": "top_p",
"temperature": 1,
"top_p": 0.95,
"top_k": 20,
"min_p": 0,
"presence_penalty": 1.5,
"repetition_penalty": 1,
"n": 1,
"max_tokens": 81920,
"skip_special_tokens": false,
"reasoning_effort": null,
"seed": 2
} | Let $D$ be the distance from school to the park in miles.
Let $P$ be Patrick's walking speed.
Let $T$ be Tanya's running speed.
Let $J$ be Jose's bicycling speed.
We are given the following relationships between their speeds:
1. Tanya ran at a constant rate of 2 miles per hour faster than Patrick walked.
So, $T = P... | stop | 186 | 8,446 | 277 | 1 | 277 | 1 | ok | evalscope-1.8.1-aime-rule-v1 | 277 | false | 277 | 1 | 1 | 1 | ok | 1 | boxed | A | correct | 0.999892 | 0.999902 | 0.000096 | 0.000002 | 0.478422 | 0.372596 | 0.148982 | Qwen/Qwen3.6-35B-A3B | 995ad96eacd98c81ed38be0c5b274b04031597b0 | 19.302777 | 19.655474 | 18.520716 | 0.209746 | 0.998478 | 3 | mean over three criterion-specific pointwise expected A--T score-token distributions; normalized to [0,1] | false | {
"prompt_sum_logprob": -470.98194338182657,
"prompt_avg_logprob": -2.545848342604468,
"prompt_ppl": 12.75404329622253,
"completion_sum_logprob": -1588.1218469332957,
"completion_avg_logprob": -0.1880324232693933,
"completion_ppl": 1.2068726454721461
} |
aime_2026 | Qwen/Qwen3.6-35B-A3B | Qwen3.6-35B-A3B | 0 | 3 | Patrick started walking at a constant rate along a straight road from school to the park. One hour after Patrick left, Tanya started running along the same road from school to the park. One hour after Tanya left, Jose started bicycling along the same road from school to the park. Tanya ran at a constant rate of $2$ mil... |
Solve the following math problem step by step. Put your answer inside \boxed{}.
{question}
Remember to put your answer inside \boxed{}. | {
"strategy": "top_p",
"temperature": 1,
"top_p": 0.95,
"top_k": 20,
"min_p": 0,
"presence_penalty": 1.5,
"repetition_penalty": 1,
"n": 1,
"max_tokens": 81920,
"skip_special_tokens": false,
"reasoning_effort": null,
"seed": 3
} | Here's a thinking process:
1. **Understand the Problem:**
- Patrick starts walking first at a constant rate. Let his speed be $v_P$.
- One hour later, Tanya starts running at a constant rate $v_T = v_P + 2$.
- One hour after Tanya leaves (so 2 hours after Patrick left), Jose starts bicycling at a constant ra... | stop | 186 | 3,563 | 277 | 1 | 277 | 1 | ok | evalscope-1.8.1-aime-rule-v1 | 277 | false | 277 | 1 | 1 | 5 | ok | 1 | boxed | A | correct | 0.999969 | 0.999975 | 0.000024 | 0.000001 | 0.548958 | 0.306338 | 0.144704 | Qwen/Qwen3.6-35B-A3B | 995ad96eacd98c81ed38be0c5b274b04031597b0 | 19.055456 | 19.819565 | 19.598634 | 0.153353 | 0.998302 | 3 | mean over three criterion-specific pointwise expected A--T score-token distributions; normalized to [0,1] | false | {
"prompt_sum_logprob": -470.98194338182657,
"prompt_avg_logprob": -2.545848342604468,
"prompt_ppl": 12.75404329622253,
"completion_sum_logprob": -375.97354525228826,
"completion_avg_logprob": -0.1055216237025788,
"completion_ppl": 1.1112901344703894
} |
aime_2026 | Qwen/Qwen3.6-35B-A3B | Qwen3.6-35B-A3B | 0 | 4 | Patrick started walking at a constant rate along a straight road from school to the park. One hour after Patrick left, Tanya started running along the same road from school to the park. One hour after Tanya left, Jose started bicycling along the same road from school to the park. Tanya ran at a constant rate of $2$ mil... |
Solve the following math problem step by step. Put your answer inside \boxed{}.
{question}
Remember to put your answer inside \boxed{}. | {
"strategy": "top_p",
"temperature": 1,
"top_p": 0.95,
"top_k": 20,
"min_p": 0,
"presence_penalty": 1.5,
"repetition_penalty": 1,
"n": 1,
"max_tokens": 81920,
"skip_special_tokens": false,
"reasoning_effort": null,
"seed": 4
} | Here's a thinking process to solve the problem:
1. **Understand the Goal:** The objective is to find the distance from the school to the park, expressed as a fraction $\frac{m}{n}$, and then calculate $m+n$.
2. **Identify the Variables and Constants:**
* **Locations:** School ($S$) and Park ($P$). Distance $d... | stop | 186 | 6,489 | 277 | 1 | 277 | 1 | ok | evalscope-1.8.1-aime-rule-v1 | 277 | false | 277 | 1 | 1 | 1 | ok | 1 | boxed | A | correct | 0.999946 | 0.999954 | 0.000045 | 0.000001 | 0.531245 | 0.350218 | 0.118537 | Qwen/Qwen3.6-35B-A3B | 995ad96eacd98c81ed38be0c5b274b04031597b0 | 19.456228 | 19.44978 | 19.813829 | 0.129694 | 0.99779 | 3 | mean over three criterion-specific pointwise expected A--T score-token distributions; normalized to [0,1] | false | {
"prompt_sum_logprob": -470.98194338182657,
"prompt_avg_logprob": -2.545848342604468,
"prompt_ppl": 12.75404329622253,
"completion_sum_logprob": -1064.8650395060035,
"completion_avg_logprob": -0.16410310363784922,
"completion_ppl": 1.1783357995374106
} |
aime_2026 | Qwen/Qwen3.6-35B-A3B | Qwen3.6-35B-A3B | 0 | 5 | Patrick started walking at a constant rate along a straight road from school to the park. One hour after Patrick left, Tanya started running along the same road from school to the park. One hour after Tanya left, Jose started bicycling along the same road from school to the park. Tanya ran at a constant rate of $2$ mil... |
Solve the following math problem step by step. Put your answer inside \boxed{}.
{question}
Remember to put your answer inside \boxed{}. | {
"strategy": "top_p",
"temperature": 1,
"top_p": 0.95,
"top_k": 20,
"min_p": 0,
"presence_penalty": 1.5,
"repetition_penalty": 1,
"n": 1,
"max_tokens": 81920,
"skip_special_tokens": false,
"reasoning_effort": null,
"seed": 5
} | Let $d$ be the distance from the school to the park in miles.
Let $v_P$ be Patrick's walking rate in miles per hour.
Let $v_T$ be Tanya's running rate in miles per hour.
Let $v_J$ be Jose's bicycling rate in miles per hour.
Let $t_P$ be the time Patrick took to reach the park.
Let $t_T$ be the time Tanya took to reach... | stop | 186 | 13,866 | 277 | 1 | 277 | 1 | ok | evalscope-1.8.1-aime-rule-v1 | 277 | false | 277 | 1 | 1 | 1 | ok | 1 | boxed | A | correct | 0.999906 | 0.999914 | 0.000085 | 0.000001 | 0.445927 | 0.445927 | 0.108147 | Qwen/Qwen3.6-35B-A3B | 995ad96eacd98c81ed38be0c5b274b04031597b0 | 17.404558 | 19.689516 | 16.388924 | 0.389624 | 0.997381 | 3 | mean over three criterion-specific pointwise expected A--T score-token distributions; normalized to [0,1] | false | {
"prompt_sum_logprob": -470.98194338182657,
"prompt_avg_logprob": -2.545848342604468,
"prompt_ppl": 12.75404329622253,
"completion_sum_logprob": -3121.825799742245,
"completion_avg_logprob": -0.22514249240893158,
"completion_ppl": 1.2525011753866926
} |
aime_2026 | Qwen/Qwen3.6-35B-A3B | Qwen3.6-35B-A3B | 0 | 6 | Patrick started walking at a constant rate along a straight road from school to the park. One hour after Patrick left, Tanya started running along the same road from school to the park. One hour after Tanya left, Jose started bicycling along the same road from school to the park. Tanya ran at a constant rate of $2$ mil... |
Solve the following math problem step by step. Put your answer inside \boxed{}.
{question}
Remember to put your answer inside \boxed{}. | {
"strategy": "top_p",
"temperature": 1,
"top_p": 0.95,
"top_k": 20,
"min_p": 0,
"presence_penalty": 1.5,
"repetition_penalty": 1,
"n": 1,
"max_tokens": 81920,
"skip_special_tokens": false,
"reasoning_effort": null,
"seed": 6
} | Let $d$ be the distance from the school to the park in miles.
Let $v_P$, $v_T$, and $v_J$ be the speeds of Patrick, Tanya, and Jose, respectively.
Let $t_P$, $t_T$, and $t_J$ be the time taken by Patrick, Tanya, and Jose to reach the park, respectively.
The problem states that they start at different times but arrive ... | stop | 186 | 12,278 | 277 | 1 | 277 | 1 | ok | evalscope-1.8.1-aime-rule-v1 | 277 | false | 277 | 1 | 1 | 3 | ok | 1 | boxed | A | correct | 0.99978 | 0.999794 | 0.000203 | 0.000003 | 0.377087 | 0.48419 | 0.138723 | Qwen/Qwen3.6-35B-A3B | 995ad96eacd98c81ed38be0c5b274b04031597b0 | 17.176286 | 19.328116 | 17.575119 | 0.396967 | 0.996986 | 3 | mean over three criterion-specific pointwise expected A--T score-token distributions; normalized to [0,1] | false | {
"prompt_sum_logprob": -470.98194338182657,
"prompt_avg_logprob": -2.545848342604468,
"prompt_ppl": 12.75404329622253,
"completion_sum_logprob": -2791.719092295778,
"completion_avg_logprob": -0.2273757201739516,
"completion_ppl": 1.2553014214170715
} |
aime_2026 | Qwen/Qwen3.6-35B-A3B | Qwen3.6-35B-A3B | 0 | 7 | Patrick started walking at a constant rate along a straight road from school to the park. One hour after Patrick left, Tanya started running along the same road from school to the park. One hour after Tanya left, Jose started bicycling along the same road from school to the park. Tanya ran at a constant rate of $2$ mil... |
Solve the following math problem step by step. Put your answer inside \boxed{}.
{question}
Remember to put your answer inside \boxed{}. | {
"strategy": "top_p",
"temperature": 1,
"top_p": 0.95,
"top_k": 20,
"min_p": 0,
"presence_penalty": 1.5,
"repetition_penalty": 1,
"n": 1,
"max_tokens": 81920,
"skip_special_tokens": false,
"reasoning_effort": null,
"seed": 7
} | Here is a thinking process to solve the problem:
1. **Understand the Goal:** The objective is to find the distance from school to the park, expressed as a fraction $m/n$, and then calculate $m+n$.
2. **Identify the Variables and Given Information:**
* Path: Straight road from School ($S$) to Park ($P$). Dista... | stop | 186 | 6,447 | 277 | 1 | 277 | 1 | ok | evalscope-1.8.1-aime-rule-v1 | 277 | false | 277 | 1 | 1 | 1 | ok | 1 | boxed | A | correct | 0.999944 | 0.999953 | 0.000045 | 0.000001 | 0.499242 | 0.357723 | 0.143035 | Qwen/Qwen3.6-35B-A3B | 995ad96eacd98c81ed38be0c5b274b04031597b0 | 19.331191 | 19.430897 | 18.270543 | 0.233432 | 0.991629 | 3 | mean over three criterion-specific pointwise expected A--T score-token distributions; normalized to [0,1] | false | {
"prompt_sum_logprob": -470.98194338182657,
"prompt_avg_logprob": -2.545848342604468,
"prompt_ppl": 12.75404329622253,
"completion_sum_logprob": -888.5191973783641,
"completion_avg_logprob": -0.13781901619022244,
"completion_ppl": 1.147767804085474
} |
aime_2026 | Qwen/Qwen3.6-35B-A3B | Qwen3.6-35B-A3B | 0 | 8 | "Patrick started walking at a constant rate along a straight road from school to the park. One hour (...TRUNCATED) | "\nSolve the following math problem step by step. Put your answer inside \\boxed{}.\n\n{question}\n\(...TRUNCATED) | {"strategy":"top_p","temperature":1.0,"top_p":0.95,"top_k":20,"min_p":0.0,"presence_penalty":1.5,"re(...TRUNCATED) | "Here is a thinking process to solve the problem:\n\n1. **Understand the Goal:** The objective is t(...TRUNCATED) | stop | 186 | 7,039 | 277 | 1 | 277 | 1 | ok | evalscope-1.8.1-aime-rule-v1 | 277 | false | 277 | 1 | 1 | 1 | ok | 1 | boxed | A | correct | 0.999894 | 0.999902 | 0.000096 | 0.000002 | 0.407368 | 0.44277 | 0.149862 | Qwen/Qwen3.6-35B-A3B | 995ad96eacd98c81ed38be0c5b274b04031597b0 | 18.272433 | 19.465021 | 19.353396 | 0.229031 | 0.99593 | 3 | "mean over three criterion-specific pointwise expected A--T score-token distributions; normalized to(...TRUNCATED) | false | {"prompt_sum_logprob":-470.98194338182657,"prompt_avg_logprob":-2.545848342604468,"prompt_ppl":12.75(...TRUNCATED) |
aime_2026 | Qwen/Qwen3.6-35B-A3B | Qwen3.6-35B-A3B | 0 | 9 | "Patrick started walking at a constant rate along a straight road from school to the park. One hour (...TRUNCATED) | "\nSolve the following math problem step by step. Put your answer inside \\boxed{}.\n\n{question}\n\(...TRUNCATED) | {"strategy":"top_p","temperature":1.0,"top_p":0.95,"top_k":20,"min_p":0.0,"presence_penalty":1.5,"re(...TRUNCATED) | "Here is a thinking process:\n\n1. **Understand the Goal**: The problem asks for the distance from (...TRUNCATED) | stop | 186 | 4,115 | 277 | 1 | 277 | 1 | ok | evalscope-1.8.1-aime-rule-v1 | 277 | false | 277 | 1 | 1 | 2 | ok | 1 | boxed | A | correct | 0.99996 | 0.999968 | 0.000031 | 0.000001 | 0.592982 | 0.280105 | 0.126913 | Qwen/Qwen3.6-35B-A3B | 995ad96eacd98c81ed38be0c5b274b04031597b0 | 19.575949 | 19.760702 | 19.634127 | 0.111477 | 0.99636 | 3 | "mean over three criterion-specific pointwise expected A--T score-token distributions; normalized to(...TRUNCATED) | false | {"prompt_sum_logprob":-470.98194338182657,"prompt_avg_logprob":-2.545848342604468,"prompt_ppl":12.75(...TRUNCATED) |
Anonymous Reasoning Lite
This repository contains data accompanying an anonymous TMLR submission. It provides 1,211,520 sampled reasoning attempts from four model configurations on competition-math and field-balanced multiple-choice questions. The data omits top-20 alternative-token distributions while retaining realized-token log probabilities and ranks.
Contents
The same attempts are available in full and metadata-only representations:
| Configuration group | Rows | Approximate size | Contents |
|---|---|---|---|
meta-math |
59,520 | 0.44 GiB | Four models and five math splits without token arrays |
meta-gpqa |
1,152,000 | 1.45 GiB | Four models on multiple-choice questions without token arrays |
Four <model>-math configs |
14,880 each | 6.28 GiB total | Math rows with realized-token arrays |
Four <model>-gpqa configs |
288,000 each | 27.05 GiB total | Multiple-choice rows with realized-token arrays |
math/<model>/<task>.parquet
gpqa/<model>/super_gpqa-NNNNN.parquet
meta-math/<model>/<task>.parquet
meta-gpqa/<model>/super_gpqa-NNNNN.parquet
One Parquet row group contains one question's 80 attempts, ordered by seed.
Loading
pip install -U datasets
from datasets import load_dataset
math_meta = load_dataset(
"AnonymizedTMLRSubmission/reasoning-lite",
"meta-math",
split="aime_2026",
streaming=True,
)
gpqa_meta = load_dataset(
"AnonymizedTMLRSubmission/reasoning-lite",
"meta-gpqa",
split="super_gpqa",
streaming=True,
)
model_data = load_dataset(
"AnonymizedTMLRSubmission/reasoning-lite",
"gpt-oss-20b_medium-math",
split="cmimc_2025",
streaming=True,
)
Data Format
| Field | Description |
|---|---|
task |
Benchmark or split identifier |
model |
Model identifier |
model_key |
Unique configuration key |
data_id, seed |
Question identifier and sampling seed |
prompt, trigger |
Problem text and generation instruction |
sampling |
Sampling configuration |
text, finish_reason |
Generated response and termination reason |
num_prompt_tokens, num_completion_tokens |
Token counts |
ground_truth |
Reference answer |
evalscope_extracted_answer, evalscope_is_correct |
Extracted answer and rule-based correctness |
cv3b_* |
Three-way verifier outputs and diagnostic values |
llmv_* |
Reference-free verifier criterion scores |
tokens |
Aggregate token statistics and, where present, token-level arrays |
Math rows additionally include boxed-answer parsing fields. Multiple-choice rows include
stage, full_data_id, uuid, selection_hash, and field taxonomy columns.
cv3b_label is the emitted A, B, or C verifier label. cv3b_prob is the full-vocabulary
probability of that emitted label. cv3b_abc_A/B/C is normalized over the three labels.
cv3b_ctx_A/B/C contains null-prompt-calibrated diagnostics and should not be interpreted
as correctness probabilities.
The reference-free verifier scores problem understanding, reasoning validity, and
conclusion support from 1 to 20. Convert an expected criterion score to the retained reward
scale with (expected - 1) / 19; average the three criterion rewards for a pointwise score.
The tokens struct contains six aggregate log-probability statistics. Per-model configs
also include realized-token, log-probability, and rank arrays. Metadata-only configs omit
the arrays.
Reproducibility
Every question has 80 attempts. Use model_key rather than model to distinguish the
three reasoning levels of the shared checkpoint. State filtering of aborted or truncated
attempts and the verifier score used in any answer-selection analysis.
License
The artifact is provided under the MIT License. Questions and problem statements from the underlying benchmarks remain subject to their applicable terms. The SuperGPQA question content is provided under ODC-BY 1.0 and requires upstream credit.
Upstream Attribution
The multiple-choice questions are derived from the SuperGPQA benchmark, published by the M-A-P Team. Users should credit that benchmark and its paper, "SuperGPQA: Scaling LLM Evaluation across 285 Graduate Disciplines" (2025), when using the question content.
Citation
Citation information will be added after anonymous review.
- Downloads last month
- -