tytodd/qwen3.5-2b-v1-embed

---
dataset_info:
- config_name: aes2_essay_scoring
features:
- name: input
struct:
- name: full_text
dtype: string
- name: prediction
struct:
- name: score
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 708937797
num_examples: 10000
download_size: 238852355
dataset_size: 708937797
- config_name: civil_comments
features:
- name: input
struct:
- name: comment
dtype: string
- name: prediction
struct:
- name: toxicity_label
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 549300287
num_examples: 10000
download_size: 179479176
dataset_size: 549300287
- config_name: colbert_humor_detection
features:
- name: input
struct:
- name: text
dtype: string
- name: prediction
struct:
- name: humor_label
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 585327253
num_examples: 10000
download_size: 190426281
dataset_size: 585327253
- config_name: customer_support_tickets_en
features:
- name: input
struct:
- name: body
dtype: string
- name: subject
dtype: string
- name: prediction
struct:
- name: queue
dtype: string
- name: type
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 358240099
num_examples: 5570
download_size: 119041172
dataset_size: 358240099
- config_name: customer_support_tickets_gorkem
features:
- name: input
struct:
- name: ticket_text
dtype: string
- name: prediction
struct:
- name: ticket_subject
dtype: string
- name: ticket_type
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 454203262
num_examples: 6775
download_size: 141359168
dataset_size: 454203262
- config_name: go_emotions
features:
- name: input
struct:
- name: text
dtype: string
- name: prediction
struct:
- name: labels
list: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 616068946
num_examples: 10000
download_size: 207413083
dataset_size: 616068946
- config_name: hh_rlhf
features:
- name: input
struct:
- name: question
dtype: string
- name: response_A
dtype: string
- name: response_B
dtype: string
- name: prediction
struct:
- name: label
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 710036037
num_examples: 10000
download_size: 243356437
dataset_size: 710036037
- config_name: lex_glue_case_hold
features:
- name: input
struct:
- name: context
dtype: string
- name: option_a
dtype: string
- name: option_b
dtype: string
- name: option_c
dtype: string
- name: option_d
dtype: string
- name: option_e
dtype: string
- name: prediction
struct:
- name: selected_option
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 668362673
num_examples: 10000
download_size: 229357085
dataset_size: 668362673
- config_name: lex_glue_scotus
features:
- name: input
struct:
- name: opinion_text
dtype: string
- name: prediction
struct:
- name: issue_id
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 552042815
num_examples: 5000
download_size: 242811361
dataset_size: 552042815
- config_name: medical_abstracts
features:
- name: input
struct:
- name: medical_abstract
dtype: string
- name: prediction
struct:
- name: condition_label
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 616044887
num_examples: 10000
download_size: 202858815
dataset_size: 616044887
- config_name: mfrc
features:
- name: input
struct:
- name: text
dtype: string
- name: prediction
struct:
- name: authority
dtype: bool
- name: care
dtype: bool
- name: fairness
dtype: bool
- name: loyalty
dtype: bool
- name: non_moral
dtype: bool
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 563017967
num_examples: 10000
download_size: 112724259
dataset_size: 563017967
- config_name: or_bench_80k
features:
- name: input
struct:
- name: prompt
dtype: string
- name: prediction
struct:
- name: or_bench_category
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 759705855
num_examples: 10000
download_size: 257389711
dataset_size: 759705855
- config_name: or_bench_hard_1k
features:
- name: input
struct:
- name: prompt
dtype: string
- name: prediction
struct:
- name: or_bench_category
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 65726176
num_examples: 1055
download_size: 22287745
dataset_size: 65726176
- config_name: projudgebench
features:
- name: input
struct:
- name: correct_answer
dtype: string
- name: question
dtype: string
- name: step_to_evaluate
dtype: string
- name: steps
list: string
- name: prediction
struct:
- name: correct
dtype: bool
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 224474033
num_examples: 2160
download_size: 67445971
dataset_size: 224474033
- config_name: reward_bench_2
features:
- name: input
struct:
- name: prompt
dtype: string
- name: response_A
dtype: string
- name: response_B
dtype: string
- name: prediction
struct:
- name: label
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 104166157
num_examples: 1492
download_size: 36234201
dataset_size: 104166157
- config_name: seekbench
features:
- name: input
struct:
- name: current_trace
dtype: string
- name: previous_traces
dtype: string
- name: question
dtype: string
- name: prediction
struct:
- name: groundness
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 30632574
num_examples: 446
download_size: 10024088
dataset_size: 30632574
- config_name: seekbench_evidence
features:
- name: input
struct:
- name: current_trace
dtype: string
- name: previous_traces
dtype: string
- name: question
dtype: string
- name: prediction
struct:
- name: clear
dtype: string
- name: sufficient
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 25402307
num_examples: 324
download_size: 8179125
dataset_size: 25402307
- config_name: seekbench_full_trace
features:
- name: input
struct:
- name: final_answer
dtype: string
- name: question
dtype: string
- name: trace
dtype: string
- name: prediction
struct:
- name: correctness
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 9609644
num_examples: 133
download_size: 3204725
dataset_size: 9609644
- config_name: sem_eval_2010_task_8
features:
- name: input
struct:
- name: sentence
dtype: string
- name: prediction
struct:
- name: relation_label
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 647254615
num_examples: 8000
download_size: 209594830
dataset_size: 647254615
- config_name: smollm_corpus
features:
- name: input
struct:
- name: text
dtype: string
- name: prediction
struct:
- name: audience
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 732580580
num_examples: 10000
download_size: 276978335
dataset_size: 732580580
- config_name: snli
features:
- name: input
struct:
- name: hypothesis
dtype: string
- name: premise
dtype: string
- name: prediction
struct:
- name: label
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 599685230
num_examples: 10000
download_size: 194419563
dataset_size: 599685230
- config_name: support_tickets_alpha
features:
- name: input
struct:
- name: description
dtype: string
- name: subject
dtype: string
- name: prediction
struct:
- name: key_phrase
dtype: string
- name: support_class
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 45439259
num_examples: 813
download_size: 14081039
dataset_size: 45439259
- config_name: toxigen_data
features:
- name: input
struct:
- name: text
dtype: string
- name: prediction
struct:
- name: toxicity_label
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 521250834
num_examples: 8960
download_size: 163382745
dataset_size: 521250834
- config_name: tweet_eval_emotion
features:
- name: input
struct:
- name: tweet
dtype: string
- name: prediction
struct:
- name: emotion_label
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 180194448
num_examples: 3257
download_size: 57392634
dataset_size: 180194448
- config_name: tweet_eval_hate
features:
- name: input
struct:
- name: tweet
dtype: string
- name: prediction
struct:
- name: hate_label
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 517425716
num_examples: 8993
download_size: 162507844
dataset_size: 517425716
- config_name: tweet_eval_irony
features:
- name: input
struct:
- name: tweet
dtype: string
- name: prediction
struct:
- name: irony_label
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 161288109
num_examples: 2862
download_size: 53776718
dataset_size: 161288109
- config_name: tweet_eval_offensive
features:
- name: input
struct:
- name: tweet
dtype: string
- name: prediction
struct:
- name: offensive_label
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 563247173
num_examples: 10000
download_size: 175693000
dataset_size: 563247173
- config_name: tweet_eval_sentiment
features:
- name: input
struct:
- name: tweet
dtype: string
- name: prediction
struct:
- name: sentiment_label
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 566581584
num_examples: 10000
download_size: 178553945
dataset_size: 566581584
- config_name: tweet_eval_stance_abortion
features:
- name: input
struct:
- name: topic
dtype: string
- name: tweet
dtype: string
- name: prediction
struct:
- name: stance_label
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 38241910
num_examples: 587
download_size: 12663569
dataset_size: 38241910
- config_name: tweet_eval_stance_atheism
features:
- name: input
struct:
- name: topic
dtype: string
- name: tweet
dtype: string
- name: prediction
struct:
- name: stance_label
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 30063423
num_examples: 461
download_size: 10249610
dataset_size: 30063423
- config_name: tweet_eval_stance_climate
features:
- name: input
struct:
- name: topic
dtype: string
- name: tweet
dtype: string
- name: prediction
struct:
- name: stance_label
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 23333936
num_examples: 355
download_size: 7727825
dataset_size: 23333936
- config_name: tweet_eval_stance_feminist
features:
- name: input
struct:
- name: topic
dtype: string
- name: tweet
dtype: string
- name: prediction
struct:
- name: stance_label
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 38429683
num_examples: 597
download_size: 12942806
dataset_size: 38429683
- config_name: tweet_eval_stance_hillary
features:
- name: input
struct:
- name: topic
dtype: string
- name: tweet
dtype: string
- name: prediction
struct:
- name: stance_label
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 37705230
num_examples: 620
download_size: 12277581
dataset_size: 37705230
- config_name: ultrafeedback
features:
- name: input
struct:
- name: prompt
dtype: string
- name: response
dtype: string
- name: prediction
struct:
- name: instruction_following
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 632776499
num_examples: 10000
download_size: 207267926
dataset_size: 632776499
- config_name: yelp
features:
- name: input
struct:
- name: text
dtype: string
- name: prediction
struct:
- name: rating
dtype: string
- name: reasoning
dtype: string
- name: messages
struct:
- name: messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: outputs
struct:
- name: reasoning_content
dtype: string
- name: text
dtype: string
- name: correct
dtype: bool
- name: embedding
list: float64
splits:
- name: train
num_bytes: 581859693
num_examples: 10000
download_size: 192231416
dataset_size: 581859693
configs:
- config_name: aes2_essay_scoring
data_files:
- split: train
path: aes2_essay_scoring/train-*
- config_name: civil_comments
data_files:
- split: train
path: civil_comments/train-*
- config_name: colbert_humor_detection
data_files:
- split: train
path: colbert_humor_detection/train-*
- config_name: customer_support_tickets_en
data_files:
- split: train
path: customer_support_tickets_en/train-*
- config_name: customer_support_tickets_gorkem
data_files:
- split: train
path: customer_support_tickets_gorkem/train-*
- config_name: go_emotions
data_files:
- split: train
path: go_emotions/train-*
- config_name: hh_rlhf
data_files:
- split: train
path: hh_rlhf/train-*
- config_name: lex_glue_case_hold
data_files:
- split: train
path: lex_glue_case_hold/train-*
- config_name: lex_glue_scotus
data_files:
- split: train
path: lex_glue_scotus/train-*
- config_name: medical_abstracts
data_files:
- split: train
path: medical_abstracts/train-*
- config_name: mfrc
data_files:
- split: train
path: mfrc/train-*
- config_name: or_bench_80k
data_files:
- split: train
path: or_bench_80k/train-*
- config_name: or_bench_hard_1k
data_files:
- split: train
path: or_bench_hard_1k/train-*
- config_name: projudgebench
data_files:
- split: train
path: projudgebench/train-*
- config_name: reward_bench_2
data_files:
- split: train
path: reward_bench_2/train-*
- config_name: seekbench
data_files:
- split: train
path: seekbench/train-*
- config_name: seekbench_evidence
data_files:
- split: train
path: seekbench_evidence/train-*
- config_name: seekbench_full_trace
data_files:
- split: train
path: seekbench_full_trace/train-*
- config_name: sem_eval_2010_task_8
data_files:
- split: train
path: sem_eval_2010_task_8/train-*
- config_name: smollm_corpus
data_files:
- split: train
path: smollm_corpus/train-*
- config_name: snli
data_files:
- split: train
path: snli/train-*
- config_name: support_tickets_alpha
data_files:
- split: train
path: support_tickets_alpha/train-*
- config_name: toxigen_data
data_files:
- split: train
path: toxigen_data/train-*
- config_name: tweet_eval_emotion
data_files:
- split: train
path: tweet_eval_emotion/train-*
- config_name: tweet_eval_hate
data_files:
- split: train
path: tweet_eval_hate/train-*
- config_name: tweet_eval_irony
data_files:
- split: train
path: tweet_eval_irony/train-*
- config_name: tweet_eval_offensive
data_files:
- split: train
path: tweet_eval_offensive/train-*
- config_name: tweet_eval_sentiment
data_files:
- split: train
path: tweet_eval_sentiment/train-*
- config_name: tweet_eval_stance_abortion
data_files:
- split: train
path: tweet_eval_stance_abortion/train-*
- config_name: tweet_eval_stance_atheism
data_files:
- split: train
path: tweet_eval_stance_atheism/train-*
- config_name: tweet_eval_stance_climate
data_files:
- split: train
path: tweet_eval_stance_climate/train-*
- config_name: tweet_eval_stance_feminist
data_files:
- split: train
path: tweet_eval_stance_feminist/train-*
- config_name: tweet_eval_stance_hillary
data_files:
- split: train
path: tweet_eval_stance_hillary/train-*
- config_name: ultrafeedback
data_files:
- split: train
path: ultrafeedback/train-*
- config_name: yelp
data_files:
- split: train
path: yelp/train-*
---