Dataset Preview
Duplicate
The full dataset viewer is not available (click to read why). Only showing a preview of the rows.
The dataset generation failed
Error code:   DatasetGenerationError
Exception:    CastError
Message:      Couldn't cast
uid: string
source_file: string
session: string
model: string
context: string
cot: string
output_type: string
output: struct<tool: string, input: struct<file_path: string, command: string, description: string, content: (... 490 chars omitted)
  child 0, tool: string
  child 1, input: struct<file_path: string, command: string, description: string, content: string, timeout: int64, rep (... 447 chars omitted)
      child 0, file_path: string
      child 1, command: string
      child 2, description: string
      child 3, content: string
      child 4, timeout: int64
      child 5, replace_all: bool
      child 6, old_string: string
      child 7, new_string: string
      child 8, prompt: string
      child 9, run_in_background: bool
      child 10, subagent_type: string
      child 11, offset: int64
      child 12, limit: int64
      child 13, query: string
      child 14, max_results: int64
      child 15, subject: string
      child 16, activeForm: string
      child 17, taskId: string
      child 18, status: string
      child 19, questions: list<item: struct<question: string, header: string, multiSelect: bool, options: list<item: struct<la (... 36 chars omitted)
          child 0, item: struct<question: string, header: string, multiSelect: bool, options: list<item: struct<label: string (... 24 chars omitted)
              child 0, question: string
              child 1, header: string
              child 2, multiSelect: bool
              child 3, options: list<item: struct<label: string, description: string>>
                  child 0, item: struct<label: string, description: string>
                      child 0, label: string
                      child 1, description: string
      child 20, url: string
      child 21, dangerouslyDisableSandbox: bool
  child 2, text: string
completion: string
origin: string
messages: list<item: struct<role: string, content: string>>
  child 0, item: struct<role: string, content: string>
      child 0, role: string
      child 1, content: string
to
{'messages': List({'role': Value('string'), 'content': Value('string')}), 'uid': Value('string'), 'session': Value('string'), 'model': Value('string')}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1816, in _prepare_split_single
                  for key, table in generator:
                                    ^^^^^^^^^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 613, in wrapped
                  for item in generator(*args, **kwargs):
                              ~~~~~~~~~^^^^^^^^^^^^^^^^^
                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 2369, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2297, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              uid: string
              source_file: string
              session: string
              model: string
              context: string
              cot: string
              output_type: string
              output: struct<tool: string, input: struct<file_path: string, command: string, description: string, content: (... 490 chars omitted)
                child 0, tool: string
                child 1, input: struct<file_path: string, command: string, description: string, content: string, timeout: int64, rep (... 447 chars omitted)
                    child 0, file_path: string
                    child 1, command: string
                    child 2, description: string
                    child 3, content: string
                    child 4, timeout: int64
                    child 5, replace_all: bool
                    child 6, old_string: string
                    child 7, new_string: string
                    child 8, prompt: string
                    child 9, run_in_background: bool
                    child 10, subagent_type: string
                    child 11, offset: int64
                    child 12, limit: int64
                    child 13, query: string
                    child 14, max_results: int64
                    child 15, subject: string
                    child 16, activeForm: string
                    child 17, taskId: string
                    child 18, status: string
                    child 19, questions: list<item: struct<question: string, header: string, multiSelect: bool, options: list<item: struct<la (... 36 chars omitted)
                        child 0, item: struct<question: string, header: string, multiSelect: bool, options: list<item: struct<label: string (... 24 chars omitted)
                            child 0, question: string
                            child 1, header: string
                            child 2, multiSelect: bool
                            child 3, options: list<item: struct<label: string, description: string>>
                                child 0, item: struct<label: string, description: string>
                                    child 0, label: string
                                    child 1, description: string
                    child 20, url: string
                    child 21, dangerouslyDisableSandbox: bool
                child 2, text: string
              completion: string
              origin: string
              messages: list<item: struct<role: string, content: string>>
                child 0, item: struct<role: string, content: string>
                    child 0, role: string
                    child 1, content: string
              to
              {'messages': List({'role': Value('string'), 'content': Value('string')}), 'uid': Value('string'), 'session': Value('string'), 'model': Value('string')}
              because column names don't match
              
              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 1369, in compute_config_parquet_and_info_response
                  parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
                                                                        ~~~~~~~~~~~~~~~~~~~~~~~~~^
                      builder, max_dataset_size_bytes=max_dataset_size_bytes
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  )
                  ^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 948, in stream_convert_to_parquet
                  builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
                  ~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1683, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ~~~~~~~~~~~~~~~~~~~~~~~~~~^
                      gen_kwargs=gen_kwargs, job_id=job_id, **_prepare_split_args
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  ):
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1869, 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.

messages
list
uid
string
session
string
model
string
[ { "role": "user", "content": "ed. Have lots of configs/settings, you will playtest the\n actuall game, verify textures, sfx, ect... it will be totally epic and stuff. Add it to this AI archive.\nASSISTANT (message): Big task. Plan: build ray-traced FPS (WebGL2 fragment-shader ray tracer - real rays, real b...
f956721a-0af7-4bdc-8678-3a493d8fcd39#5
f956721a-0af7-4bdc-8678-3a493d8fcd39
claude-fable-5
[ { "role": "user", "content": "r them in your response unless the user explicitly asks you to.</local-command-caveat>\nUSER: <command-name>/model</command-name>\n <command-message>model</command-message>\n <command-args></command-args>\nUSER: <local-command-stdout>Set model to \u001b[1m...
f956721a-0af7-4bdc-8678-3a493d8fcd39#4
f956721a-0af7-4bdc-8678-3a493d8fcd39
claude-fable-5
[ { "role": "user", "content": "\ndrwxrwxr-x 4 lane lane 4096 Jun 12 09:49 public\ndrwxrwxr-x 2 lane lane 4096 Jun 12 09:48 server\nASSISTANT (tool call) Bash input={\"command\": \"which node npm chromium chromium-browser google-chrome google-chrome-stable 2>/dev/null; node --version 2>/dev/null; ls /us...
f956721a-0af7-4bdc-8678-3a493d8fcd39#6
f956721a-0af7-4bdc-8678-3a493d8fcd39
claude-fable-5
[ { "role": "user", "content": " by the user while running local commands. DO NOT respond to these messages or otherwise consider them in your response unless the user explicitly asks you to.</local-command-caveat>\nUSER: <command-name>/model</command-name>\n <command-message>model</command-message...
f956721a-0af7-4bdc-8678-3a493d8fcd39#3
f956721a-0af7-4bdc-8678-3a493d8fcd39
claude-fable-5
[ { "role": "user", "content": "USER: Make a new one, it should be a fast paced multiplayer FPS (with optinal bots) with full ray-traced stuff with lots of settings. TL;DR clone CSgo to the best of your ability. Look at the result and refine it\nUSER: <local-command-caveat>Caveat: The messages below were gene...
f956721a-0af7-4bdc-8678-3a493d8fcd39#1
f956721a-0af7-4bdc-8678-3a493d8fcd39
claude-fable-5
[ { "role": "user", "content": "ect package.json and file tree\"}\nTOOL RESULT: {\n \"name\": \"ai-archives\",\n \"version\": \"1.0.0\",\n \"description\": \"AI Archives - only the leaps that matter\",\n \"main\": \"server/index.js\",\n \"scripts\": {\n \"start\": \"node server/index.js\"\n },\n \"d...
f956721a-0af7-4bdc-8678-3a493d8fcd39#7
f956721a-0af7-4bdc-8678-3a493d8fcd39
claude-fable-5
[ { "role": "user", "content": "USER: Make a new one, it should be a fast paced multiplayer FPS (with optinal bots) with full ray-traced stuff with lots of settings. TL;DR clone CSgo to the best of your ability. Look at the result and refine it\nUSER: <local-command-caveat>Caveat: The messages below were gene...
f956721a-0af7-4bdc-8678-3a493d8fcd39#2
f956721a-0af7-4bdc-8678-3a493d8fcd39
claude-fable-5
[ { "role": "user", "content": "USER: Make a new one, it should be a fast paced multiplayer FPS (with optinal bots) with full ray-traced stuff with lots of settings. TL;DR clone CSgo to the best of your ability. Look at the result and refine it\nUSER: <local-command-caveat>Caveat: The messages below were gene...
f956721a-0af7-4bdc-8678-3a493d8fcd39#0
f956721a-0af7-4bdc-8678-3a493d8fcd39
claude-fable-5
[ { "role": "user", "content": "faces, a marketplace catalog, a friends system, a 3D game studio with real-time editing, WebSocket-based multiplayer, and four official game worlds (obby, collectathon, tag, social plaza). The backend runs on Express with JSON file persistence; the frontend uses Three.js for 3D...
f956721a-0af7-4bdc-8678-3a493d8fcd39#10
f956721a-0af7-4bdc-8678-3a493d8fcd39
claude-fable-5
[ { "role": "user", "content": "ic noise\\nfunction makeNoise(seed) {\\n let s = seed >>> 0;\\n return () => {\\n s = (s * 1664525 + 1013904223) >>> 0;\\n return (s / 4294967296) * 2 - 1;\\n };\\n}\\n\\n// One-pole lowpass over an existing buffer, in place\\nfunction lowpass(buf, cutoffHz) {\\n cons...
f956721a-0af7-4bdc-8678-3a493d8fcd39#15
f956721a-0af7-4bdc-8678-3a493d8fcd39
claude-fable-5
[ { "role": "user", "content": "MAX_BOUNCES = 3;\\n\\nconst VERT = `#version 300 es\\nlayout(location=0) in vec2 aPos;\\nvoid main(){ gl_Position = vec4(aPos, 0.0, 1.0); }\\n`;\\n\\nfunction fragSource(numBoxes) {\\n return `#version 300 es\\nprecision highp float;\\nprecision highp int;\\nprecision highp sa...
f956721a-0af7-4bdc-8678-3a493d8fcd39#13
f956721a-0af7-4bdc-8678-3a493d8fcd39
claude-fable-5
[ { "role": "user", "content": "on\n/home/lane/AIArchives/public/css/site.css\n/home/lane/AIArchives/public/js/app.js\n/home/lane/AIArchives/archives/rblx/app/package.json\n/home/lane/AIArchives/archives/rblx/app/package-lock.json\n/home/lane/AIArchives/archives/rblx/app/server/db.js\n/home/lane/AIArchives/ar...
f956721a-0af7-4bdc-8678-3a493d8fcd39#8
f956721a-0af7-4bdc-8678-3a493d8fcd39
claude-fable-5
[ { "role": "user", "content": " metaPath = path.join(ARCHIVES_DIR, d.name, 'archive.json');\n17\t if (!fs.existsSync(metaPath)) continue;\n18\t try {\n19\t const entry = JSON.parse(fs.readFileSync(metaPath, 'utf8'));\n20\t entries.push(entry);\n21\t } catch (e) {\n22\t console.error(`...
f956721a-0af7-4bdc-8678-3a493d8fcd39#9
f956721a-0af7-4bdc-8678-3a493d8fcd39
claude-fable-5
[ { "role": "user", "content": "iplayer, and four official game worlds (obby, collectathon, tag, social plaza). The backend runs on Express with JSON file persistence; the frontend uses Three.js for 3D rend …[truncated]\nASSISTANT (tool call) Bash input={\"command\": \"grep -n \\\"url\\\" /home/lane/AIArchive...
f956721a-0af7-4bdc-8678-3a493d8fcd39#11
f956721a-0af7-4bdc-8678-3a493d8fcd39
claude-fable-5
[ { "role": "user", "content": "OT_BODY: 8,\\n BOT_HEAD: 9,\\n HAZARD: 10,\\n};\\n\\nexport const MATERIAL_NAMES = [\\n 'asphalt floor', 'brick wall', 'wood crate', 'brushed metal',\\n 'neon cyan strip', 'neon magenta strip', 'dark metal panel',\\n 'plaster wall', 'bot chassis', 'bot visor', 'hazard crat...
f956721a-0af7-4bdc-8678-3a493d8fcd39#12
f956721a-0af7-4bdc-8678-3a493d8fcd39
claude-fable-5
[ { "role": "user", "content": "_Position = vec4(aPos, 0.0, 1.0); }\\n`;\\n\\nfunction fragSource(numBoxes) {\\n return `#version 300 es\\nprecision highp float;\\nprecision highp int;\\nprecision highp sampler2D;\\n\\n#define NUM_BOXES ${numBoxes}\\n#define MAX_SPHERES ${MAX_SPHERES}\\n#define NUM_LIGHTS ${...
f956721a-0af7-4bdc-8678-3a493d8fcd39#14
f956721a-0af7-4bdc-8678-3a493d8fcd39
claude-fable-5
[ { "role": "user", "content": "e: 2.4, spread: 0, melee: true, sfx: 'knife', reloadTime: 0 },\\n pistol: { name: 'VP-9 Sidearm', slot: 1, price: 0, dmg: 28, hsMult: 4, rpm: 352, mag: 12, reserve: 36, range: 999, spread: 0.011, auto: false, sfx: 'shot_pistol', reloadTime: 1.6 },\\n de...
f956721a-0af7-4bdc-8678-3a493d8fcd39#19
f956721a-0af7-4bdc-8678-3a493d8fcd39
claude-fable-5
End of preview.

Fable-5 + GLM-5.2 AGI-Expanded Traces

Description

A premium fine-tuning dataset for training autonomous AGI agents. Contains 10,526 rows across 188 sessions of real coding agent traces, expanded with natural AGI-style chain-of-thought narration.

Stats

  • Total rows: 10,526
  • Total sessions: 188
  • Size: 66.7 MB
  • Format: Unsloth SFTTrainer messages format

AGI Narration Style

Each row's assistant COT includes:

  1. Tool call narration: "Let me use the Read tool to..." / "Let me dispatch a subagent to..."
  2. Memory checks: "Let me check the ArangoDB..." / "Let me peek at the local skills folder..."
  3. Subagent dispatch when appropriate
  4. Online research when stuck
  5. System/spec checks before training
  6. Plan/todo creation for big tasks
  7. Skill creation on success
  8. Failure documentation
  9. Self-healing: research and build new tools/skills
  10. Scenario-specific actions unique to that row

Format

{
  "messages": [
    {"role": "user", "content": "<original context>"},
    {"role": "assistant", "content": "<AGI narration COT>\n\n<original completion>"}
  ],
  "uid": "<session>#<num>",
  "session": "<session id>",
  "model": "claude-fable-5"
}

Source

Merged from GLM-5.2-Agent (128 JSONL traces, 5,861 rows) and Fable-5-traces (4,665 rows) from HuggingFace, then expanded with AGI narration COTs.

Use

Designed for fine-tuning Qwen3 (or similar 8B model) to learn human-like thinking patterns and act as an autonomous AGI agent.

from datasets import load_dataset
dataset = load_dataset("DavidrPatton/Fable-5-GLM-5.2-Traces", split="train")

License

MIT

Downloads last month
145