Datasets:

KarthikRagunathAnandaKumar
/

sliderl-multi-turn-rollouts

Error code:   DatasetGenerationError
Exception:    TypeError
Message:      Couldn't cast array of type struct<slide_index: int64, title: string, bullet_points: list<item: string>, sections_per_slide: int64> to null
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1890, in _prepare_split_single
                  writer.write_table(table)
                File "/usr/local/lib/python3.12/site-packages/datasets/arrow_writer.py", line 760, in write_table
                  pa_table = table_cast(pa_table, self._schema)
                             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2272, in table_cast
                  return cast_table_to_schema(table, schema)
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2224, in cast_table_to_schema
                  cast_array_to_feature(
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 1795, in wrapper
                  return pa.chunked_array([func(chunk, *args, **kwargs) for chunk in array.chunks])
                                           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2052, in cast_array_to_feature
                  casted_array_values = _c(array.values, feature.feature)
                                        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 1797, in wrapper
                  return func(array, *args, **kwargs)
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2086, in cast_array_to_feature
                  return array_cast(
                         ^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 1797, in wrapper
                  return func(array, *args, **kwargs)
                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 1950, in array_cast
                  raise TypeError(f"Couldn't cast array of type {_short_str(array.type)} to {_short_str(pa_type)}")
              TypeError: Couldn't cast array of type struct<slide_index: int64, title: string, bullet_points: list<item: string>, sections_per_slide: int64> to null
              
              The above exception was the direct cause of the following exception:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1347, in compute_config_parquet_and_info_response
                  parquet_operations = convert_to_parquet(builder)
                                       ^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 980, in convert_to_parquet
                  builder.download_and_prepare(
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 884, in download_and_prepare
                  self._download_and_prepare(
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 947, in _download_and_prepare
                  self._prepare_split(split_generator, **prepare_split_kwargs)
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1739, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/builder.py", line 1922, in _prepare_split_single
                  raise DatasetGenerationError("An error occurred while generating the dataset") from e
              datasets.exceptions.DatasetGenerationError: An error occurred while generating the dataset

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

episode_id string	model string	brief string	brief_topic string	turns list	total_steps int64	final_phase string	completed bool	slides_created int64	cumulative_reward float64	final_quality dict	outline list	slides_html list	rollout_idx int64	elapsed_seconds float64
79ad8d29-63fb-4d3b-bc0f-aca9656406ae	sliderl-base	{"topic":"2026 Marketing Budget Allocation Strategy","audience":"CMO and marketing leadership","num_slides":8,"sections_per_slide":3,"confidence":1.0,"content":{"total_budget":"$8.4M (up 12% from 2025)","allocation":{"digital_advertising":{"amount":"$2.8M","percent":"33%"},"content_marketing":{"amount":"$1.4M","percent...	2026 Marketing Budget Allocation Strategy	[ { "turn": 0, "assistant": "```json\n{\n \"tool\": \"set_theme\",\n \"theme_name\": \"creative\"\n}\n```", "tool_call": "{\"tool\":\"set_theme\",\"theme_name\":\"creative\"}", "observation": "Tool result (success=True):\nTheme set to 'creative'.\n\nState: phase=RESEARCH, slides=0/8, turns remaining...	12	DONE	true	8	1.018651	{ "code_rules": 0.9796325607, "render_quality": 1, "claude_aesthetic_html": 0.705, "claude_aesthetic_visual": 0.60625, "content_quality": 0.9, "brief_reconstruction": 0.67, "aggregate": 0.8086507561 }	[]	[ "<!DOCTYPE html>\n<html>\n<head>\n<meta charset=\"utf-8\">\n<style>\n @import url('https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600;700&display=swap');\n \n * { margin: 0; padding: 0; box-sizing: border-box; }\n \n body {\n width: 1280px;\n height: 720px;\n font-fa...	0	5
c5d0aaf8-568a-4c8a-8e67-5d912c0c8ae5	sliderl-base	{"topic":"5G Technology Business Impact Analysis","audience":"telecom executives","num_slides":7,"sections_per_slide":3,"confidence":0.3,"preferred_theme":"creative"}	5G Technology Business Impact Analysis	[ { "turn": 0, "assistant": "```json\n{\n \"tool\": \"web_search\",\n \"query\": \"5G technology market analysis, current trends, key players, ROI projections\"\n}\n```", "tool_call": "{\"tool\":\"web_search\",\"query\":\"5G technology market analysis, current trends, key players, ROI projections\"}", ...	17	DONE	true	7	1.0181	{ "code_rules": 0.9631443733, "render_quality": 1, "claude_aesthetic_html": 0.7285714286, "claude_aesthetic_visual": 0.53, "content_quality": 0.85, "brief_reconstruction": 0.665, "aggregate": 0.7881001516 }	[]	[ "<!DOCTYPE html>\n<html>\n<head>\n<meta charset=\"utf-8\">\n<style>\n @import url('https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600;700&display=swap');\n \n * { margin: 0; padding: 0; box-sizing: border-box; }\n \n body {\n width: 1280px;\n height: 720px;\n font-fa...	1	5.5
4219448f-7507-4717-bfed-249258890bb9	sliderl-base	"{\"topic\":\"Agile Transformation Success Metrics\",\"audience\":\"engineering leadership\",\"num_s(...TRUNCATED)	Agile Transformation Success Metrics	[{"turn":0,"assistant":"```json\n{\n \"tool\": \"set_theme\",\n \"theme_name\": \"tech\"\n}\n```",(...TRUNCATED)	33	DONE	true	6	1.178456	{"code_rules":0.9545550897,"render_quality":1.0,"claude_aesthetic_html":0.5733333333,"claude_aesthet(...TRUNCATED)	[]	["<!DOCTYPE html>\n<html>\n<head>\n<meta charset=\"utf-8\">\n<style>\n @import url('https://fonts(...TRUNCATED)	2	7.4
d7813ad5-f54d-4a52-a20d-c213c42b9673	sliderl-base	"{\"topic\":\"API Economy and Monetization Strategies\",\"audience\":\"product leadership\",\"num_sl(...TRUNCATED)	API Economy and Monetization Strategies	[{"turn":0,"assistant":"```json\n{\n \"tool\": \"web_search\",\n \"query\": \"API economy trends, (...TRUNCATED)	35	GENERATE	false	7	0.439077	{"code_rules":0.9650549451,"render_quality":1.0,"claude_aesthetic_html":0.7257142857,"claude_aesthet(...TRUNCATED)	[]	["<!DOCTYPE html>\n<html>\n<head>\n<meta charset=\"utf-8\">\n<style>\n @import url('https://fonts(...TRUNCATED)	3	5
adfea10d-1d98-47d2-8f21-c186ee22102d	sliderl-base	"{\"topic\":\"Artificial Intelligence in Manufacturing\",\"audience\":\"manufacturing executives\",\(...TRUNCATED)	Artificial Intelligence in Manufacturing	[{"turn":0,"assistant":"```json\n{\n \"tool\": \"web_search\",\n \"query\": \"Artificial Intellige(...TRUNCATED)	35	GENERATE	false	7	0.399497	{"code_rules":0.903302485,"render_quality":1.0,"claude_aesthetic_html":0.7542857143,"claude_aestheti(...TRUNCATED)	[]	["<!DOCTYPE html>\n<html>\n<head>\n<meta charset=\"utf-8\">\n<style>\n @import url('https://fonts(...TRUNCATED)	4	5
9d8cc98f-4112-4365-b63a-6e17698f984e	sliderl-base	"{\"topic\":\"Autonomous Vehicle Technology Progress\",\"audience\":\"automotive investors\",\"num_s(...TRUNCATED)	Autonomous Vehicle Technology Progress	[{"turn":0,"assistant":"```json\n{\n \"tool\": \"web_search\",\n \"query\": \"autonomous vehicle t(...TRUNCATED)	35	REFINE	false	8	0.965625	{"code_rules":0.906875,"render_quality":1.0,"claude_aesthetic_html":0.5875,"claude_aesthetic_visual"(...TRUNCATED)	[]	["<!DOCTYPE html>\n<html>\n<head>\n<meta charset=\"utf-8\">\n<style>\n @import url('https://fonts(...TRUNCATED)	5	156.7
bac940f4-17f0-41eb-a023-f44fc81c5bdb	sliderl-base	"{\"topic\":\"B2B Sales Automation Technology Stack\",\"audience\":\"sales operations\",\"num_slides(...TRUNCATED)	B2B Sales Automation Technology Stack	[{"turn":0,"assistant":"```json\n{\n \"action\": \"set_theme\",\n \"theme_name\": \"creative\"\n}\(...TRUNCATED)	35		false	0	0	{"code_rules":0.0,"render_quality":0.0,"claude_aesthetic_html":0.0,"claude_aesthetic_visual":0.0,"co(...TRUNCATED)	[]	[]	6	5
bbd418f4-e575-4fc2-ba34-5eada2adc0d3	sliderl-base	"{\"topic\":\"Blockchain Enterprise Use Cases 2026\",\"audience\":\"technology leadership\",\"num_sl(...TRUNCATED)	Blockchain Enterprise Use Cases 2026	[{"turn":0,"assistant":"```json\n{\n \"tool\": \"web_search\",\n \"query\": \"Blockchain enterpris(...TRUNCATED)	35	GENERATE	false	7	0.394906	{"code_rules":0.9719131445,"render_quality":1.0,"claude_aesthetic_html":0.7357142857,"claude_aesthet(...TRUNCATED)	[]	["<!DOCTYPE html>\n<html>\n<head>\n<meta charset=\"utf-8\">\n<style>\n @import url('https://fonts(...TRUNCATED)	7	5
c9e64ae3-a418-4e0f-bbb8-f1ca89ef7cf7	sliderl-base	"{\"topic\":\"Cloud Computing Market Share Analysis - AWS vs Azure vs GCP\",\"audience\":\"IT strate(...TRUNCATED)	Cloud Computing Market Share Analysis - AWS vs Azure vs GCP	[{"turn":0,"assistant":"```json\n{\n \"tool\": \"set_theme\",\n \"theme_name\": \"creative\"\n}\n`(...TRUNCATED)	30	DONE	true	8	1.035857	{"code_rules":0.9619052823,"render_quality":1.0,"claude_aesthetic_html":0.71875,"claude_aesthetic_vi(...TRUNCATED)	[]	[]	8	119.7
78ebf869-4ae9-4a14-9062-97328e126ca1	sliderl-base	"{\"topic\":\"Cloud Cost Optimization Strategies\",\"audience\":\"CFO and IT leadership\",\"num_slid(...TRUNCATED)	Cloud Cost Optimization Strategies	[{"turn":0,"assistant":"```json\n{\n \"action\": \"web_search\",\n \"query\": \"Cloud cost optimiz(...TRUNCATED)	35		false	0	0	{"code_rules":0.0,"render_quality":0.0,"claude_aesthetic_html":0.0,"claude_aesthetic_visual":0.0,"co(...TRUNCATED)	[]	[]	9	5

End of preview.

SlideRL Multi-Turn Rollouts

Multi-turn rollout trajectories from the SlideRL environment across 6 different models. Each trajectory captures an agent interacting with the SlideRL slide-generation environment over multiple tool-use turns.

Models

File	Model	Episodes	Description
`claude_opus_4_6_trajectories.json`	Claude Opus 4.6	48	Anthropic Claude Opus 4.6
`claude_sonnet_4_6_trajectories.json`	Claude Sonnet 4.6	48	Anthropic Claude Sonnet 4.6
`llama_4_scout_trajectories.json`	Llama 4 Scout	48	Meta Llama 4 Scout
`gpt_oss_120b_trajectories.json`	GPT-OSS-120B	48	OpenAI GPT-OSS-120B
`finetuned_model_trajectories.json`	SlideRL Finetuned	48	GRPO-finetuned model
`base_model_trajectories.json`	SlideRL Base	48	Base model before finetuning

Trajectory Format

Each JSON file contains a list of episodes. Each episode has:

{
  "episode_id": "uuid",
  "model": "model-name",
  "brief": { "topic": "...", "audience": "...", "num_slides": N, ... },
  "brief_topic": "Topic Name",
  "turns": [
    {
      "turn": 0,
      "assistant": "model's response text",
      "tool_call": { "tool": "tool_name", ... },
      "observation": "environment feedback",
      "success": true,
      "phase": "RESEARCH|PLAN|BUILD|REFINE",
      "slide_count": 0,
      "done": false,
      "step_reward": 0.01,
      "cumulative_reward": 0.01
    }
  ],
  "total_steps": 25,
  "final_phase": "REFINE",
  "completed": true,
  "slides_created": 6,
  "cumulative_reward": 1.15,
  "final_quality": { "aggregate": 0.80, ... },
  "slides_html": ["<slide html>..."],
  "elapsed_seconds": 120.5
}

Environment

The SlideRL environment is a multi-turn tool-use environment where agents create presentation slide decks by:

Researching the topic via web search
Planning an outline
Building slides with HTML/CSS
Refining slides for quality

Agents interact through structured tool calls (web_search, create_outline, set_theme, add_slide, replace_slide, delete_slide, etc.) and receive environment observations with quality feedback.

48 Diverse Business Topics

All models are evaluated on the same 48 business presentation briefs spanning finance, technology, marketing, cybersecurity, healthcare, and more.

Citation

Part of the SlideRL project — an open-ended RL environment for training LLM agents on multi-turn tool-use tasks.

Downloads last month: 48

Total file size:

21 MB

Models trained or fine-tuned on KarthikRagunathAnandaKumar/sliderl-multi-turn-rollouts

KarthikRagunathAnandaKumar/LearningToPresent-RL-Qwen-2.5B-Coder-Instruct-GRPO-Finetuned

Text Generation • Updated Mar 18 • 3