datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
Qwen3-0.6B-pts
Qwen/Qwen3-0.6B — Pivotal Token Search
Pivotal reasoning events for Qwen/Qwen3-0.6B, at three representational scales in one
file, produced with PTS.
latent meta-token / workspace event (Latent PTS) ← J-lens readout
↓
emitted pivotal token (Token PTS) ← Phi-4 PTS
↓
sentence-level thought anchor (Sentence PTS) ← Thought Anchors
↓
success / failure probability shift
All three are CausalReasoningEvent records — one schema… See the full description on the dataset page: https://huggingface.co/datasets/codelion/Qwen3-0.6B-pts.trinity-coordinator-adapted-qwen3-0.6bQwen3-0.6B-pts-steering-vectors
PTS Steering Vectors Dataset
A dataset of activation-based steering vectors created using the Pivotal Token Search (PTS) technique.
Details
Source: Generated using the PTS tool
Model: Qwen/Qwen3-0.6B
Dataset Structure
This dataset contains:
steering_vectors.jsonl: The main file with token-level steering vectors
Usage
These steering vectors can be used for activation-based steering during inference to guide language models toward particular… See the full description on the dataset page: https://huggingface.co/datasets/codelion/Qwen3-0.6B-pts-steering-vectors.trinity-coordinator-adapted-qwen3-0.6bQwen3-0.6B-pts-thought-anchors
PTS Thought Anchors Dataset
A dataset of thought anchors - critical reasoning steps - identified using the Thought Anchors technique from the PTS tool.
Details
Source: Generated using the PTS tool
Model: Qwen/Qwen3-0.6B
Tags: pts, thought-anchors, reasoning, llm-analysis
Dataset Structure
This dataset contains thought anchors identified from reasoning traces. Each anchor represents a sentence that significantly impacts the success probability of the reasoning… See the full description on the dataset page: https://huggingface.co/datasets/codelion/Qwen3-0.6B-pts-thought-anchors.Qwen3-0.6B-pts-dpo-pairs
PTS DPO Dataset
A Direct Preference Optimization (DPO) dataset created using the Pivotal Token Search (PTS) technique.
Details
Source: Generated using the PTS tool
Model: Qwen/Qwen3-0.6B
Format
Each example in the dataset consists of:
prompt: The context leading up to the pivotal token
chosen: The preferred token that increases success probability
rejected: The alternative token that decreases success probability
metadata: Additional information about the… See the full description on the dataset page: https://huggingface.co/datasets/codelion/Qwen3-0.6B-pts-dpo-pairs.massive_serve_dpr_wiki_qwen3_0.6b_ivfpqQwen3-0.6B-icmcountdown-qwen3-0.6b
Countdown Qwen3-0.6B Pass@10 Buckets
Countdown arithmetic problems filtered by observed local Qwen/Qwen3-0.6B success rate over 10 rollouts per problem.
Each problem asks for an arithmetic expression that reaches a target using each listed source number at most once. The final answer should be inside \boxed{...}. Canonical solutions are provided, but any verifier-valid expression is accepted.
Subsets
subset
source bucket
count
observed successes out of 10… See the full description on the dataset page: https://huggingface.co/datasets/simpissa/countdown-qwen3-0.6b.Qwen3-0.6B-icm-dpo-pairsqwen3-0.6B-interleaved-thinking-data
Qwen3 0.6B Interleaved Thinking Data
This dataset contains 8,704 pretraining-style text chunks augmented with short interleaved teacher thoughts. It was built for the blog post Self-Improving Pretraining as a Substrate for Agentic Post-Training.
The dataset turns ordinary pretraining text into the supervised stage of a thinking mid-training pipeline. A teacher inserts short local thoughts into raw FineWeb-Edu chunks while preserving the original text. The student then learns the… See the full description on the dataset page: https://huggingface.co/datasets/Jarrodbarnes/qwen3-0.6B-interleaved-thinking-data.qwen3-0.6b-blind-spots
Qwen3-0.6B-Base Blind Spots Dataset
Dataset Description
This dataset contains 12 diverse examples of failure cases ("blind spots") identified in the Qwen3-0.6B-Base model, a pretrained base language model released in May 2025 with 0.6 billion parameters.
Model Information
Model: Qwen/Qwen3-0.6B-Base
Parameters: 0.6B (600 million)
Type: Base model (pretrained, not instruction-tuned)
Context Length: 32,768 tokens
Release Date: May 2025
Architecture:… See the full description on the dataset page: https://huggingface.co/datasets/AmmarHashme/qwen3-0.6b-blind-spots.gsm8k-onpolicy-Qwen3-0.6B-cptqwen3_0.6b_rstarcoder_instill_n8_valredundancy5_round1qwen3_0.6b_openthoughts3_math53K_instill_n8_valredundancy5_round1qwen3_0.6B_configThinkSafe-0.6Bgsm8k-onpolicy-Qwen3-0.6B-iftqwen3_0.6b_openthoughts4_science26K_instill_n8_valredundancy5_round1qwen3_0.6b_openthoughts4_code9K_instill_n8_valredundancy5_round1ThinkSafe-0.6B-n1radiology-index-qwen3-embedding-0.6bgsm8k-onpolicy-Qwen3-0.6B-dpogsm8k-onpolicy-Qwen3-0.6B-dpo-dedupedqwen-0.6B_20000_5ZaryaOrthrusDataset-0.6BThinkSafe-0.6B-v3Refusal Statistics:Total harmful prompts: 17888Number of refusals: 2722Number of non-refusals: 15166Refusal rate: 15.22%
Filtering Statistics:Total examples before filtering: 39997Total examples after filtering: 39096Examples removed (harmful): 901Safe ratio: 97.75%
FINAL SUMMARY
Harmful prompts that model refused: 2722Harmful prompts with generated refusals: 15166Benign prompts: 22112Total examples after LlamaGuard… See the full description on the dataset page: https://huggingface.co/datasets/Seanie-lee/ThinkSafe-0.6B-v3.ThinkSafe-0.6B-n5ThinkSafe-0.6B-s1lora-rules-qwen3-0.6b-r8-n180
LoRA Rules Dataset
Synthetic behavioral rules dataset for training a hypernetwork that generates
LoRA adapters on-the-fly from structured rule strings.
Format
Each record is a JSON line with fields:
rule_id — unique identifier
rule_type — one of: Constraint, Format, Knowledge, Persona, Safety, Tone
weight — float 0.0–1.0, importance of the rule
description — natural language rule description
raw — full rule string [RuleType|Weight] Description
training_examples — list of… See the full description on the dataset page: https://huggingface.co/datasets/broadfield-dev/lora-rules-qwen3-0.6b-r8-n180.