--- pretty_name: "TRACE — Taxonomy-Referenced ABA Clinical Examples" license: cc-by-nc-4.0 language: - en size_categories: - 1K"}, {"role": "user", "content": ""}, {"role": "assistant", "content": ""} ], "meta": { "task_type": "teaching_program" | "session_interpretation", "example_id": "", "gold_labels": { ... task-specific labels ... }, "provenance": { "layer": 1, "area": "dtt" | "net" | "task_analysis" | "session_interpretation", "template_id": "...", "taxonomy_cells": { ... exact sampled values ... }, "teacher_model": null, "seed_tag": "...", "generated_at": "..." } } } ``` Only `messages` is used for training (mask_prompt: true). `meta` is preserved for evaluation and provenance tracking. ### 3.2 Gold labels per task **Teaching program:** ``` { "method": "dtt" | "net" | "task_analysis", "domain": "VB-MAPP." | "AFLS.", "level": "L1" | "L2" | "L3" | ..., # VB-MAPP only "learner_profile": "early" | "school_age" | "adolescent" | "adult", "mastery_state": "emerging" | "developing" | "approaching" | "near" | "mastered" | "generalization", "program_type": "independence" | "toleration", # Task Analysis only "chain_type": "forward" | "backward" | "total_task" # Task Analysis only } ``` **Session interpretation:** ``` { "pattern_class": "", "behavior_functions": { "": "escape"|"attention"|"tangible"|"automatic"|"unknown" }, "escalation_level": 1 | 2 | 3 | 4, "confidence": "high" | "moderate" | "low", "crisis_plan_required": true | false } ``` ### 3.3 Provenance Every sampled choice is recorded in `meta.provenance.taxonomy_cells`. For teaching programs this includes the skill target, prompt hierarchy, reinforcement schedule, error correction, and mastery criterion. For session interpretation it includes the pattern, number of sessions, behaviors sampled, functions inferred, and whether IOA and ABC data are present. This enables: - **Traceability:** any clinical issue in a generated example can be traced back to the exact taxonomy cell that produced it. - **Reproducibility:** the dataset can be regenerated deterministically from the configs and seed. - **Stratification:** splits can be regenerated with different stratification keys without regenerating the corpus. - **Ablation:** subsets can be constructed by filtering on provenance cells. --- ## 4. How It Was Built ### 4.1 Architecture ``` configs/ ├── shared/ # cross-area primitives (learner profiles, mastery states, prompt types) ├── dtt/ # DTT area: taxonomy + template + compatibility ├── net/ # NET area ├── task_analysis/ # Chaining area (independence + toleration templates) └── session_interpretation/ # Session interp area: # taxonomy + compatibility + trajectory_rules # + recommendations (per-pattern bullets) + template src/generators/ ├── aba_dtt.py # per-area generator ├── aba_net.py ├── aba_task_analysis.py └── aba_session_interp.py src/generate.py # orchestrator ``` ### 4.2 Generation loop For each example: 1. Sample a clinical configuration from the taxonomy, weighted by realistic clinical frequency (level weights, module weights, pattern weights). 2. Apply compatibility rules (e.g., DTT errorless error correction pairs only with most-to-least prompting; certain patterns have bias toward certain behavior functions). 3. Compute template slots — current-prompt-level guidance from mastery state; MO arrangement from skill keywords; mastery criterion; reinforcement schedule. 4. Render `user_content` from a random user-variant. 5. Render `assistant_content` by filling `{slots}` in the assistant template. 6. Stamp `meta.gold_labels` and `meta.provenance`. 7. Write as one JSONL line. ### 4.3 Grounding - **Skill curricula** — VB-MAPP milestones (Sundberg 2008) for verbal-behavior domains; AFLS (Partington & Mueller) for adaptive-living domains. - **Teaching methods** — DTT (Lovaas 1987; Smith 2001); NET (Hart & Risley 1975 in CHH Ch. 18); Task Analysis / chaining (CHH Ch. 20). - **Operational definitions of target behaviors** — Cooper/Heron/Heward Ch. 3, 27; key JABA papers (Iwata et al. 1994 for SIB and functional analysis; Carr & Durand 1985 for FCT; Hanley, Iwata, & McCord 2003 for FBA). - **Session patterns** — derived from CHH Ch. 6–7 (analyzing behavior change) and Stokes & Baer 1977 (generalization). - **Crisis plans** — BACB Ethics Code (2020) §3.05; ABAI Position Statement on Restraint and Seclusion (2010). Physical-intervention procedures are left vague on purpose because they vary by training program (Safety-Care, CPI, PMT, TCI) and jurisdiction; the dataset emphasizes verbal de-escalation, environmental safety, BIP authorization, and contraindications. ### 4.4 Clinical-accuracy pipeline Initial generation produced the structural skeleton. The corpus was then iterated via **practitioner-in-the-loop review**: a full-text render of every held-out candidate was browsed by a reviewer with ABA practitioner exposure, flags were raised for clinically inaccurate topographies, and each flag triggered a taxonomy or rendering fix plus a regeneration and a commit. Examples of iterations shipped: | Issue flagged | Fix | |---|---| | "Context: R/D" tag leaking a specific facility's tool | Removed from all generators and templates | | "Raises hand to request" wrong topography for mand | Replaced with AAC, vocal, or full-sentence requesting; added Tolerating Denied Access as a replacement-behavior target | | "Whining during trials" in frustration indicators | Removed; replaced with property destruction, aggression, peer aggression, SIB onset | | Pica measured as simple frequency | Split into attempts / successful / unsuccessful | | Toleration programs missing entirely | Added as a distinct program type for hygiene routines | | Crisis plans too generic | Rewritten with tiered framework, BIP-authorization requirement, contraindications, BACB Ethics citations | | `maintenance` mastery state (post-mastery retention) appeared as a current teaching state | Removed; maintenance is retained only as a post-mastery probe concept in generalization plans | | Only one dishwashing variant in Home Skills | Added by-hand, operating dishwasher, unloading dishwasher | | Token-economy backups abstract | Made concrete for children (candy, cookies, leisure, screen time, sensory item) | | Fecal smearing and toileting accidents absent | Added as target behaviors with attempts/intercepted/completed and urine/BM in-toilet/accidents rendering | | Bathroom requesting absent from mand curriculum | Added to NET L1 mands with bathroom_opportunity MO | Each iteration is one commit — the provenance chain is preserved end-to-end. --- ## 5. Intended Uses ### 5.1 Direct use - Instruction-tuning a small language model (recommended: Gemma 4 E2B 4-bit with QLoRA) to draft ABA teaching programs and interpret session logs. - Evaluation of task-specific competencies using the held-out `test.jsonl` (279 examples). - Research on taxonomy-driven synthetic data generation for clinical decision support. ### 5.2 Downstream use - Pilot **drafting assistants** for BCBAs, always requiring clinician review before use with learners. - Comparison baselines for future ABA-specific LLM work (Kumar et al. 2024 "Personalized-ABA" is the closest direct predecessor; the present dataset extends to structured program generation and session-log interpretation). ### 5.3 Out-of-scope (do NOT use for) - Autonomous clinical decisions. - Writing final Behavior Intervention Plans without BCBA review. - Training models on or combined with real client records (the pipeline and schema are explicitly designed to avoid this). - Legal or insurance documentation. - Medical diagnosis. --- ## 6. Ethics & Risks ### 6.1 Provenance Every example is synthetic. No real client data, no real session notes, no real identifiers were used at any step. Learner references use synthetic IDs (`SYN-####`); dates fall in the range 2026-01-01 to 2026-12-31. ### 6.2 Clinical-risk framing The dataset is framed — in system prompts, in the dataset card, and in the paper — as producing **drafts for clinician review**. Assistant responses are structured so that a BCBA can quickly see the method, the stimulus arrangement, the prompt hierarchy, the reinforcement plan, the error-correction procedure, the mastery criterion, and the generalization plan; approve what is clinically appropriate; and revise what is not. Session-interpretation responses surface a confidence level (high / moderate / low) and an escalation level (1–4) explicitly so the clinician can calibrate action. ### 6.3 Crisis-plan sensitivity Crisis plans were written against the ABAI 2010 Position Statement on Restraint and Seclusion and BACB Ethics Code 2020 §3.05. The dataset references facility crisis-prevention frameworks (Safety-Care, CPI, PMT, TCI) only as examples and **deliberately avoids specifying restraint procedures** because those procedures are (a) jurisdiction-dependent, (b) training-certification-gated, and (c) learner-specific (many learners have contraindications). The dataset embeds explicit text in every crisis-plan bullet that physical intervention is used only when specifically authorized in the learner's BIP and only by staff currently certified in the facility's training program. ### 6.4 Population representation Learner profiles are intentionally abstract (early / school-age / adolescent / adult). The dataset does not encode demographic categories (race, socioeconomic status, gender identity) and does not attempt to characterize clinical presentations by such categories. This is a deliberate choice for a first release; future versions may add representation if grounded in published demographic work. ### 6.5 Known limitations - **English only.** Teaching method terminology, curriculum targets, and session-log conventions are rendered in English. - **Synthetic distributions.** Pattern frequencies are approximately uniform for learnability; real clinical practice has different frequencies (mastery-progression is far more common than frustration-pattern in a healthy caseload). The dataset is explicitly a teaching set, not an epidemiological sample. - **VB-MAPP + AFLS only.** Other curricula (ABLLS-R, Essential for Living, PEAK) are not covered. Practitioners using those curricula should adapt. - **No longitudinal data.** Sessions within a log are temporally ordered but the pipeline does not model real continuity over months or years. - **Toleration covers hygiene only.** Other toleration programs (e.g., wearing glasses, riding in a car seat) are not represented. - **Toilet-training acquisition is out of scope.** Accidents are tracked in session logs and bathroom-requesting is taught as a NET mand, but the full Azrin & Foxx 1971 rapid toilet-training acquisition protocol is not included as a task-analysis program. This is a planned v2 addition. ### 6.6 License TBD — recommend CC BY-NC 4.0 (research use, non-commercial) pending the hackathon submission requirements. Clinical use of the dataset or any model fine-tuned on it requires facility approval and BCBA oversight. --- ## 7. Reproducibility The corpus is regenerated deterministically from: - `configs/` (all YAMLs) - `configs/generation.yaml` (seed and per-area counts) - `src/generators/*.py` (generator code) ```bash uv run python src/generate.py --all # regenerate the 2999-example corpus uv run python src/split_data.py # stratified split uv run python src/prepare_curation.py # browseable review.md uv run python src/compile_curation.py # test.jsonl + sanity.jsonl ``` The repo commit pinned for this dataset version is `945248d`. --- ## 8. Citation Please cite as: > Kahunla, F. (2026). *TRACE: Taxonomy-Grounded Synthetic Data for Teaching Program Generation and Session Interpretation in Applied Behavior Analysis.* Drexel University CS 614 final project / Gemma 4 Good Hackathon submission. --- ## 9. Appendices - **Datasheet** (Gebru et al. 2021 format): `docs/dataset/datasheet.md` - **Data statement** (Bender & Friedman 2018 format): `docs/dataset/data-statement.md` - **Taxonomy reference** (operational definitions + citations): `docs/dataset/taxonomy-v1.md` - **Schema reference** (wire format + slot specifications): `docs/dataset/schema-v1.md` - **Research paper draft:** `docs/paper/paper-draft.md`