inside-out-replication-canary-v1

Canary run: 5 questions per relation, 50 samples, Llama-3-8B. Full pipeline E2E test.

Dataset Info

• Rows: 482
• Columns: 11

Columns

Column	Type	Description
relation	Value('string')	Wikidata relation (P26=spouse, P264=label, P176=manufacturer, P50=author)
question_id	Value('string')	Unique question identifier
question	Value('string')	Entity-centric question text
gold_answer	Value('string')	Ground truth answer from Wikidata
answer	Value('string')	Model-generated answer (greedy or sampled)
label	Value('string')	Judge verdict: CORRECT or INCORRECT
log_p_a_q	Value('float64')	Log probability P(answer
log_p_norm_a_q	Value('float64')	Length-normalized log probability P_norm(answer
p_true_v0	Value('float64')	Restricted softmax P(True) using A/B verification prompt (V0)
verif_v1_score	Value('float64')	Restricted softmax P(True) using True/False prompt (V1)
verif_v2_score	Value('float64')	Restricted softmax P(True) using Yes/No prompt (V2)

Generation Parameters

{
  "script_name": "run_canary.py",
  "model": "meta-llama/Meta-Llama-3-8B-Instruct",
  "description": "Canary run: 5 questions per relation, 50 samples, Llama-3-8B. Full pipeline E2E test.",
  "experiment_name": "inside-out-replication",
  "job_id": "mll:26166",
  "cluster": "mll",
  "artifact_status": "final",
  "canary": true,
  "hyperparameters": {
    "n_questions": 5,
    "n_samples": 50,
    "temperature": 1.0,
    "judge_model": "Qwen/Qwen2.5-14B-Instruct"
  },
  "input_datasets": []
}

Usage

from datasets import load_dataset

dataset = load_dataset("latkes/inside-out-replication-canary-v1", split="train")
print(f"Loaded {len(dataset)} rows")

---