Dataset Viewer
The dataset viewer is not available for this split.
Cannot load the dataset split (in streaming mode) to extract the first rows.
Error code: StreamingRowsError
Exception: CastError
Message: Couldn't cast
messages: list<item: struct<role: string, content: string, reasoning_content: string>>
child 0, item: struct<role: string, content: string, reasoning_content: string>
child 0, role: string
child 1, content: string
child 2, reasoning_content: string
metadata: struct<scenario_id: string, trait_id: string, trait_name: string, trait_text: string, domain: string (... 38 chars omitted)
child 0, scenario_id: string
child 1, trait_id: string
child 2, trait_name: string
child 3, trait_text: string
child 4, domain: string
child 5, shortcut: string
child 6, situation: string
run_id: string
wall_clock_s: double
smoke: bool
workers: int64
stage_seconds: struct<refine: double, respond: double, rewrite: double>
child 0, refine: double
child 1, respond: double
child 2, rewrite: double
constitution_sha256: string
config: struct<seed: int64, constitution: string, n_traits: int64, scenarios_per_trait: int64, scenarios_per (... 560 chars omitted)
child 0, seed: int64
child 1, constitution: string
child 2, n_traits: int64
child 3, scenarios_per_trait: int64
child 4, scenarios_per_call: int64
child 5, output_dir: string
child 6, hf_repo: string
child 7, hf_repo_smoke: string
child 8, hf_private: bool
child 9, workers: int64
child 10, budget_usd: double
child 11, defaults: struct<temperature: double, max_tokens: int64>
child 0, temperature: double
child 1, max_tokens: int64
child 12, models: struct<scenarios: struct<model
...
hild 0, anthropic/claude-sonnet-5: struct<calls: int64, prompt_tokens: int64, completion_tokens: int64, usd: double>
child 0, calls: int64
child 1, prompt_tokens: int64
child 2, completion_tokens: int64
child 3, usd: double
child 1, anthropic/claude-haiku-4.5: struct<calls: int64, prompt_tokens: int64, completion_tokens: int64, usd: double>
child 0, calls: int64
child 1, prompt_tokens: int64
child 2, completion_tokens: int64
child 3, usd: double
child 1, by_stage: struct<refine: struct<calls: int64, prompt_tokens: int64, completion_tokens: int64, usd: double>, re (... 181 chars omitted)
child 0, refine: struct<calls: int64, prompt_tokens: int64, completion_tokens: int64, usd: double>
child 0, calls: int64
child 1, prompt_tokens: int64
child 2, completion_tokens: int64
child 3, usd: double
child 1, respond: struct<calls: int64, prompt_tokens: int64, completion_tokens: int64, usd: double>
child 0, calls: int64
child 1, prompt_tokens: int64
child 2, completion_tokens: int64
child 3, usd: double
child 2, rewrite: struct<calls: int64, prompt_tokens: int64, completion_tokens: int64, usd: double>
child 0, calls: int64
child 1, prompt_tokens: int64
child 2, completion_tokens: int64
child 3, usd: double
child 2, total_usd: double
run_dir: string
git_sha: string
to
{'run_id': Value('string'), 'git_sha': Value('string'), 'smoke': Value('bool'), 'constitution_sha256': Value('string'), 'config': {'seed': Value('int64'), 'constitution': Value('string'), 'n_traits': Value('int64'), 'scenarios_per_trait': Value('int64'), 'scenarios_per_call': Value('int64'), 'output_dir': Value('string'), 'hf_repo': Value('string'), 'hf_repo_smoke': Value('string'), 'hf_private': Value('bool'), 'workers': Value('int64'), 'budget_usd': Value('float64'), 'defaults': {'temperature': Value('float64'), 'max_tokens': Value('int64')}, 'models': {'scenarios': {'model': Value('string'), 'temperature': Value('float64'), 'max_tokens': Value('int64')}, 'draft': {'model': Value('string'), 'temperature': Value('float64'), 'max_tokens': Value('int64')}, 'refine': {'model': Value('string'), 'temperature': Value('float64'), 'max_tokens': Value('int64')}, 'respond': {'model': Value('string'), 'temperature': Value('float64'), 'max_tokens': Value('int64')}, 'rewrite': {'model': Value('string'), 'temperature': Value('float64'), 'max_tokens': Value('int64')}}}, 'effective': {'n_traits': Value('int64'), 'scenarios_per_trait': Value('int64'), 'scenarios_per_call': Value('int64')}, 'counts': {'traits': Value('int64'), 'scenarios': Value('int64'), 'drafts': Value('int64'), 'refined': Value('int64'), 'responses': Value('int64'), 'final': Value('int64'), 'sft': Value('int64')}, 'usage': {'by_model': {'anthropic/claude-sonnet-5': {'calls': Value('int64'), 'prompt_tokens': Value('int64'), 'completion_tokens': Value('int64'), 'usd': Value('float64')}, 'anthropic/claude-haiku-4.5': {'calls': Value('int64'), 'prompt_tokens': Value('int64'), 'completion_tokens': Value('int64'), 'usd': Value('float64')}}, 'by_stage': {'refine': {'calls': Value('int64'), 'prompt_tokens': Value('int64'), 'completion_tokens': Value('int64'), 'usd': Value('float64')}, 'respond': {'calls': Value('int64'), 'prompt_tokens': Value('int64'), 'completion_tokens': Value('int64'), 'usd': Value('float64')}, 'rewrite': {'calls': Value('int64'), 'prompt_tokens': Value('int64'), 'completion_tokens': Value('int64'), 'usd': Value('float64')}}, 'total_usd': Value('float64')}, 'wall_clock_s': Value('float64'), 'stage_seconds': {'refine': Value('float64'), 'respond': Value('float64'), 'rewrite': Value('float64')}, 'workers': Value('int64'), 'hf_repo': Value('string'), 'run_dir': Value('string')}
because column names don't match
Traceback: Traceback (most recent call last):
File "/src/services/worker/src/worker/utils.py", line 147, in get_rows_or_raise
return get_rows(
dataset=dataset,
...<4 lines>...
column_names=column_names,
)
File "/src/libs/libcommon/src/libcommon/utils.py", line 272, in decorator
return func(*args, **kwargs)
File "/src/services/worker/src/worker/utils.py", line 127, in get_rows
rows_plus_one = list(itertools.islice(safe_iter(ds, dataset=dataset), rows_max_number + 1))
File "/src/services/worker/src/worker/utils.py", line 483, in safe_iter
yield from ds.decode(False) if ds.features else ds
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2840, in __iter__
for key, example in ex_iterable:
^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2373, in __iter__
for key, pa_table in self._iter_arrow():
~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 2398, in _iter_arrow
for key, pa_table in self.ex_iterable._iter_arrow():
~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 536, in _iter_arrow
for key, pa_table in iterator:
^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/iterable_dataset.py", line 419, in _iter_arrow
for key, pa_table in self.generate_tables_fn(**gen_kwags):
~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 343, in _generate_tables
self._cast_table(pa_table, json_field_paths=json_field_paths),
~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/usr/local/lib/python3.14/site-packages/datasets/packaged_modules/json/json.py", line 132, in _cast_table
pa_table = table_cast(pa_table, self.info.features.arrow_schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2378, in table_cast
return cast_table_to_schema(table, schema)
File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2306, in cast_table_to_schema
raise CastError(
...<3 lines>...
)
datasets.table.CastError: Couldn't cast
messages: list<item: struct<role: string, content: string, reasoning_content: string>>
child 0, item: struct<role: string, content: string, reasoning_content: string>
child 0, role: string
child 1, content: string
child 2, reasoning_content: string
metadata: struct<scenario_id: string, trait_id: string, trait_name: string, trait_text: string, domain: string (... 38 chars omitted)
child 0, scenario_id: string
child 1, trait_id: string
child 2, trait_name: string
child 3, trait_text: string
child 4, domain: string
child 5, shortcut: string
child 6, situation: string
run_id: string
wall_clock_s: double
smoke: bool
workers: int64
stage_seconds: struct<refine: double, respond: double, rewrite: double>
child 0, refine: double
child 1, respond: double
child 2, rewrite: double
constitution_sha256: string
config: struct<seed: int64, constitution: string, n_traits: int64, scenarios_per_trait: int64, scenarios_per (... 560 chars omitted)
child 0, seed: int64
child 1, constitution: string
child 2, n_traits: int64
child 3, scenarios_per_trait: int64
child 4, scenarios_per_call: int64
child 5, output_dir: string
child 6, hf_repo: string
child 7, hf_repo_smoke: string
child 8, hf_private: bool
child 9, workers: int64
child 10, budget_usd: double
child 11, defaults: struct<temperature: double, max_tokens: int64>
child 0, temperature: double
child 1, max_tokens: int64
child 12, models: struct<scenarios: struct<model
...
hild 0, anthropic/claude-sonnet-5: struct<calls: int64, prompt_tokens: int64, completion_tokens: int64, usd: double>
child 0, calls: int64
child 1, prompt_tokens: int64
child 2, completion_tokens: int64
child 3, usd: double
child 1, anthropic/claude-haiku-4.5: struct<calls: int64, prompt_tokens: int64, completion_tokens: int64, usd: double>
child 0, calls: int64
child 1, prompt_tokens: int64
child 2, completion_tokens: int64
child 3, usd: double
child 1, by_stage: struct<refine: struct<calls: int64, prompt_tokens: int64, completion_tokens: int64, usd: double>, re (... 181 chars omitted)
child 0, refine: struct<calls: int64, prompt_tokens: int64, completion_tokens: int64, usd: double>
child 0, calls: int64
child 1, prompt_tokens: int64
child 2, completion_tokens: int64
child 3, usd: double
child 1, respond: struct<calls: int64, prompt_tokens: int64, completion_tokens: int64, usd: double>
child 0, calls: int64
child 1, prompt_tokens: int64
child 2, completion_tokens: int64
child 3, usd: double
child 2, rewrite: struct<calls: int64, prompt_tokens: int64, completion_tokens: int64, usd: double>
child 0, calls: int64
child 1, prompt_tokens: int64
child 2, completion_tokens: int64
child 3, usd: double
child 2, total_usd: double
run_dir: string
git_sha: string
to
{'run_id': Value('string'), 'git_sha': Value('string'), 'smoke': Value('bool'), 'constitution_sha256': Value('string'), 'config': {'seed': Value('int64'), 'constitution': Value('string'), 'n_traits': Value('int64'), 'scenarios_per_trait': Value('int64'), 'scenarios_per_call': Value('int64'), 'output_dir': Value('string'), 'hf_repo': Value('string'), 'hf_repo_smoke': Value('string'), 'hf_private': Value('bool'), 'workers': Value('int64'), 'budget_usd': Value('float64'), 'defaults': {'temperature': Value('float64'), 'max_tokens': Value('int64')}, 'models': {'scenarios': {'model': Value('string'), 'temperature': Value('float64'), 'max_tokens': Value('int64')}, 'draft': {'model': Value('string'), 'temperature': Value('float64'), 'max_tokens': Value('int64')}, 'refine': {'model': Value('string'), 'temperature': Value('float64'), 'max_tokens': Value('int64')}, 'respond': {'model': Value('string'), 'temperature': Value('float64'), 'max_tokens': Value('int64')}, 'rewrite': {'model': Value('string'), 'temperature': Value('float64'), 'max_tokens': Value('int64')}}}, 'effective': {'n_traits': Value('int64'), 'scenarios_per_trait': Value('int64'), 'scenarios_per_call': Value('int64')}, 'counts': {'traits': Value('int64'), 'scenarios': Value('int64'), 'drafts': Value('int64'), 'refined': Value('int64'), 'responses': Value('int64'), 'final': Value('int64'), 'sft': Value('int64')}, 'usage': {'by_model': {'anthropic/claude-sonnet-5': {'calls': Value('int64'), 'prompt_tokens': Value('int64'), 'completion_tokens': Value('int64'), 'usd': Value('float64')}, 'anthropic/claude-haiku-4.5': {'calls': Value('int64'), 'prompt_tokens': Value('int64'), 'completion_tokens': Value('int64'), 'usd': Value('float64')}}, 'by_stage': {'refine': {'calls': Value('int64'), 'prompt_tokens': Value('int64'), 'completion_tokens': Value('int64'), 'usd': Value('float64')}, 'respond': {'calls': Value('int64'), 'prompt_tokens': Value('int64'), 'completion_tokens': Value('int64'), 'usd': Value('float64')}, 'rewrite': {'calls': Value('int64'), 'prompt_tokens': Value('int64'), 'completion_tokens': Value('int64'), 'usd': Value('float64')}}, 'total_usd': Value('float64')}, 'wall_clock_s': Value('float64'), 'stage_seconds': {'refine': Value('float64'), 'respond': Value('float64'), 'rewrite': Value('float64')}, 'workers': Value('int64'), 'hf_repo': Value('string'), 'run_dir': Value('string')}
because column names don't matchNeed help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.
YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
synthdoc v2 — difficult-advice SFT
| field | value |
|---|---|
experiment |
Constitution-grounded synthetic difficult-advice conversations (a user facing an ethically ambiguous shortcut; the assistant reasons about its values and declines norm-violations), used as SFT data to reduce agentic misalignment on Qwen3.6. Stage 7 = final SFT split of the synthdoc_v2 pipeline. |
date_generated |
2026-08-03 |
constitution |
constitutions/claude_distilled_12_principles_mid/constitution.md (sha256 fe2ed96093d68a871fb15669e8fea9d357fb9b51f5affff15380f62ee749a642); 12-principles mid-distillation of the Claude constitution, in-repo under constitutions/ |
source_repo |
teaching_claude_why_replication @ commit 96ff8aa36dd7d9914b016f88e80b1e3724f96ccc (https://github.com/Matthew-Bozoukov/teaching_claude_why_replication) |
models |
via OpenRouter — scenarios/draft/respond: anthropic/claude-haiku-4.5; refine/rewrite: anthropic/claude-sonnet-5 |
generation_config |
seed 0; n_traits 9, scenarios_per_trait 245. Per-stage sampling — scenarios: anthropic/claude-haiku-4.5 T=1.1 max_tokens=8192; draft: anthropic/claude-haiku-4.5 T=1.0 max_tokens=2048; refine: anthropic/claude-sonnet-5 T=0.7 max_tokens=6144; respond: anthropic/claude-haiku-4.5 T=1.0 max_tokens=4096; rewrite: anthropic/claude-sonnet-5 T=0.7 max_tokens=12288 |
schema |
JSONL; 2203 rows. Each row: messages = [system, user, assistant] (system sets an agentic assistant persona/tools; user presents the ethically-ambiguous situation; assistant reasons and declines the norm-violation) and metadata = {scenario_id, trait_id, trait_name, trait_text, domain, shortcut, situation}. |
provenance |
Generated by the synthdoc_v2 / synth pipeline in the source repo at the pinned commit: uv run synth run --config configs/data/synth/difficult_advice.yaml (6-stage: traits -> scenarios -> draft prompts -> refined prompts -> draft responses -> final -> SFT). Full run manifest.json accompanies this file. Wall-clock 5096.7s; generation cost $166.72. |
- Downloads last month
- -