factprobe-replication-generation-spouse-7b-base-v1

Free-form spouse generation: for every P26 subject NAME form, the base 7B model was asked 'Who is the {spouse} of ? Answer with just the name:' with 4 in-context demos, and sampled 5 times with nucleus sampling (top_p=0.95, temperature=1.0, max_tokens=64, stop at newline). 28,815 subject names. Companion to the P(Yes) probing datasets — this is what the model GENERATES, not a yes/no score.

Dataset Info

• Rows: 27815
• Columns: 12

Columns

Column	Type	Description
relation	Value('string')	P26 (spouse).
subject	Value('string')	subject entity QID.
alias_s	Value('string')	the subject NAME string put in the prompt (surface-form level; every alias probed).
model	Value('string')	HF model id.
revision	Value('string')	HF revision (main = base).
top_p	Value('float64')	nucleus sampling top_p.
temperature	Value('float64')	sampling temperature.
n	Value('int64')	samples per subject name.
max_tokens	Value('int64')	generation cap.
prompt_hash	Value('string')	sha1[:12] of the exact prompt.
samples	List(Value('string'))	the 5 generated answers (free-form names), one nucleus sample each; full text, no truncation.
demo_leak	Value('bool')	No description provided

Generation Parameters

{
  "script_name": "generate_spouse.py",
  "model": "allenai/OLMo-2-1124-7B @ main (base, end of stage-2 mid-training)",
  "description": "Free-form spouse generation: for every P26 subject NAME form, the base 7B model was asked 'Who is the {spouse} of <name>? Answer with just the name:' with 4 in-context demos, and sampled 5 times with nucleus sampling (top_p=0.95, temperature=1.0, max_tokens=64, stop at newline). 28,815 subject names. Companion to the P(Yes) probing datasets \u2014 this is what the model GENERATES, not a yes/no score.",
  "hyperparameters": {
    "top_p": 0.95,
    "temperature": 1.0,
    "n_samples": 5,
    "max_tokens": 64,
    "relation": "P26"
  },
  "experiment_name": "factprobe-replication",
  "job_id": "mll:78619",
  "cluster": "mll",
  "artifact_status": "final",
  "canary": false,
  "input_datasets": []
}

Usage

from datasets import load_dataset

dataset = load_dataset("latkes/factprobe-replication-generation-spouse-7b-base-v1", split="train")
print(f"Loaded {len(dataset)} rows")

---