datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
lemonseed-sudoku
lemonseed-sudoku
LemonSeed — 4x4 Latin-square / Sudoku completion with stepwise solving.
Contents
sudoku.jsonl (5000 rows)
Format
JSON Lines (.jsonl), one example per line.
Provenance
Synthetic, generated programmatically for the LemonSeed 1.5B project (by Geramy L. Loveless). Data authored by Michael Anthony Falabella.
lemonseed-mixed-skills
lemonseed-mixed-skills
LemonSeed — mixed Go + Sudoku + addition + prose stream.
Contents
mixed.jsonl (11428 rows)
Format
JSON Lines (.jsonl), one example per line.
Provenance
Synthetic, generated programmatically for the LemonSeed 1.5B project (by Geramy L. Loveless). Data authored by Michael Anthony Falabella.
lemonseed-rl-tasks-v2
lemonseed-rl-tasks-v2
LemonSeed — GRPO RL tasks v2 (vocab MC / antonym MC / cloze MC / grammar / dialogue / arithmetic).
Contents
rl_tasks_v2.jsonl (12492 rows)
Format
JSON Lines (.jsonl), one example per line.
Provenance
Synthetic, generated programmatically for the LemonSeed 1.5B project (by Geramy L. Loveless). Data authored by Michael Anthony Falabella.
lemonseed-cumulative-r5
lemonseed-cumulative-r5
LemonSeed — cumulative round 5 (first fully successful multi-skill revival: prose + arithmetic + scaffold removal).
Contents
cum_r5.jsonl (20000 rows)
Format
JSON Lines (.jsonl), one example per line.
Provenance
Synthetic, generated programmatically for the LemonSeed 1.5B project (by Geramy L. Loveless). Data authored by Michael Anthony Falabella.
lemonseed-games-r2
lemonseed-games-r2
LemonSeed — games round 2 (Go/Sudoku atari + constraint reasoning).
Contents
games_r2.jsonl (12000 rows)
Format
JSON Lines (.jsonl), one example per line.
Provenance
Synthetic, generated programmatically for the LemonSeed 1.5B project (by Geramy L. Loveless). Data authored by Michael Anthony Falabella.
lemonseed-cumulative-r2
lemonseed-cumulative-r2
LemonSeed — cumulative round 2 (multi-skill with arithmetic replay).
Contents
cum_r2.jsonl (20000 rows)
Format
JSON Lines (.jsonl), one example per line.
Provenance
Synthetic, generated programmatically for the LemonSeed 1.5B project (by Geramy L. Loveless). Data authored by Michael Anthony Falabella.
lemonseed-arithmetic
lemonseed-arithmetic
LemonSeed — multi-digit arithmetic scratchpad (decomposition method).
Contents
arithmetic.jsonl (16000 rows)
Format
JSON Lines (.jsonl), one example per line.
Provenance
Synthetic, generated programmatically for the LemonSeed 1.5B project (by Geramy L. Loveless). Data authored by Michael Anthony Falabella.
lemonseed-rl-tasks-cogen-snapshot
lemonseed-rl-tasks-cogen-snapshot
LemonSeed — RL co-gen task snapshot.
Contents
rl_tasks_cogen_snapshot.jsonl (1809 rows)
Format
JSON Lines (.jsonl), one example per line.
Provenance
Synthetic, generated programmatically for the LemonSeed 1.5B project (by Geramy L. Loveless). Data authored by Michael Anthony Falabella.
qwen_qa_pairs_cli_training.jsonl
Data sources
Multiple datasets from Hugging Face related to natural language to CLI pairs were gathered.
Human reviewed synthetic data from Claude Opus4.6 and ChatGPT4.5 were added.
A handful of grounding rows related to the organisation "Spicy Lemonade" were added (see details below)
Data processing
As part of the processing, data was converted to the Alpaca format with instruction (natural language), input (typically blank) and output (the CLI command) columns.
The… See the full description on the dataset page: https://huggingface.co/datasets/spicy-lemonade/qwen_qa_pairs_cli_training.jsonl.lemonseed-cot-math-clean
lemonseed-cot-math-clean
LemonSeed — cleaned chain-of-thought math word problems.
Contents
cot_math_clean.jsonl (3530 rows)
Format
JSON Lines (.jsonl), one example per line.
Provenance
Synthetic, generated programmatically for the LemonSeed 1.5B project (by Geramy L. Loveless). Data authored by Michael Anthony Falabella.
lemonseed-rl-tasks-cogen
lemonseed-rl-tasks-cogen
LemonSeed — RL co-generated chat tasks (teacher-authored).
Contents
rl_tasks_cogen.jsonl (2500 rows)
Format
JSON Lines (.jsonl), one example per line.
Provenance
Synthetic, generated programmatically for the LemonSeed 1.5B project (by Geramy L. Loveless). Data authored by Michael Anthony Falabella.
lemonseed-rl-chat-tasks
lemonseed-rl-chat-tasks
LemonSeed — chat-alignment RL tasks (prompt/gold single-turn).
Contents
rl_chat_tasks.jsonl (9852 rows)
Format
JSON Lines (.jsonl), one example per line.
Provenance
Synthetic, generated programmatically for the LemonSeed 1.5B project (by Geramy L. Loveless). Data authored by Michael Anthony Falabella.
lemonseed-sudoku-mixed
lemonseed-sudoku-mixed
LemonSeed — Sudoku reasoning interleaved with arithmetic replay.
Contents
sudoku_mixed.jsonl (5714 rows)
Format
JSON Lines (.jsonl), one example per line.
Provenance
Synthetic, generated programmatically for the LemonSeed 1.5B project (by Geramy L. Loveless). Data authored by Michael Anthony Falabella.
lemonseed-teacher-sft
lemonseed-teacher-sft
LemonSeed — teacher-model SFT traces (multiple-choice vocabulary).
Contents
teacher_sft.jsonl (3333 rows)
Format
JSON Lines (.jsonl), one example per line.
Provenance
Synthetic, generated programmatically for the LemonSeed 1.5B project (by Geramy L. Loveless). Data authored by Michael Anthony Falabella.
lemonseed-word-problems
lemonseed-word-problems
LemonSeed — 13-category arithmetic word problems with plan + scratchpad.
Contents
wordproblems.jsonl (2500 rows)
Format
JSON Lines (.jsonl), one example per line.
Provenance
Synthetic, generated programmatically for the LemonSeed 1.5B project (by Geramy L. Loveless). Data authored by Michael Anthony Falabella.
lemonseed-vocabulary
lemonseed-vocabulary
LemonSeed — WordNet vocabulary Q&A with chain-of-thought definitions.
Contents
vocab.jsonl (4000 rows)
Format
JSON Lines (.jsonl), one example per line.
Provenance
Synthetic, generated programmatically for the LemonSeed 1.5B project (by Geramy L. Loveless). Data authored by Michael Anthony Falabella.
lemonseed-go-mixed
lemonseed-go-mixed
LemonSeed — Go reasoning interleaved with arithmetic replay.
Contents
go_mixed.jsonl (5714 rows)
Format
JSON Lines (.jsonl), one example per line.
Provenance
Synthetic, generated programmatically for the LemonSeed 1.5B project (by Geramy L. Loveless). Data authored by Michael Anthony Falabella.
lemonseed-addition-carry-method
lemonseed-addition-carry-method
LemonSeed — carry-method addition scratchpad (LSB-first, single-digit facts). Teaches digit-level addition with explicit written steps.
Contents
addition.jsonl (10000 rows)
Format
JSON Lines (.jsonl), one example per line.
Provenance
Synthetic, generated programmatically for the LemonSeed 1.5B project (by Geramy L. Loveless). Data authored by Michael Anthony Falabella.
lemonseed-go
lemonseed-go
LemonSeed — 5x5 Go liberty/atari/capture reasoning with verification steps.
Contents
go.jsonl (5000 rows)
Format
JSON Lines (.jsonl), one example per line.
Provenance
Synthetic, generated programmatically for the LemonSeed 1.5B project (by Geramy L. Loveless). Data authored by Michael Anthony Falabella.
gemma_qa_pairs_cli_training.jsonl
Data sources
Multiple datasets from Hugging Face related to natural language to CLI pairs were gathered.
Human reviewed synthetic data from Claude Opus4.6 and ChatGPT4.5 were added.
A handful of grounding rows related to the organisation "Spicy Lemonade" were added (see details below)
Data processing
As part of the processing, data was converted to the Alpaca format with instruction (natural language), input (typically blank) and output (the CLI command) columns.
The… See the full description on the dataset page: https://huggingface.co/datasets/spicy-lemonade/gemma_qa_pairs_cli_training.jsonl.lemonseed-cot-math
lemonseed-cot-math
LemonSeed — chain-of-thought math word problems (11 generators, computed solutions).
Contents
cot_math.jsonl (4000 rows)
Format
JSON Lines (.jsonl), one example per line.
Provenance
Synthetic, generated programmatically for the LemonSeed 1.5B project (by Geramy L. Loveless). Data authored by Michael Anthony Falabella.
lemonseed-cumulative-r1
lemonseed-cumulative-r1
LemonSeed — cumulative round 1 (Go 30% / Sudoku 30% / addition 28% / prose 12%).
Contents
cum_r1.jsonl (15000 rows)
Format
JSON Lines (.jsonl), one example per line.
Provenance
Synthetic, generated programmatically for the LemonSeed 1.5B project (by Geramy L. Loveless). Data authored by Michael Anthony Falabella.
lemonseed-cumulative-r4
lemonseed-cumulative-r4
LemonSeed — cumulative round 4 (multi-skill, prose recovering).
Contents
cum_r4.jsonl (20000 rows)
Format
JSON Lines (.jsonl), one example per line.
Provenance
Synthetic, generated programmatically for the LemonSeed 1.5B project (by Geramy L. Loveless). Data authored by Michael Anthony Falabella.
lemonseed-rl-tasks-cogen-pilot
lemonseed-rl-tasks-cogen-pilot
LemonSeed — RL co-gen pilot tasks.
Contents
rl_tasks_cogen_pilot.jsonl (61 rows)
Format
JSON Lines (.jsonl), one example per line.
Provenance
Synthetic, generated programmatically for the LemonSeed 1.5B project (by Geramy L. Loveless). Data authored by Michael Anthony Falabella.
lemonseed-rl-tasks
lemonseed-rl-tasks
LemonSeed — GRPO RL tasks (vocab/cloze/antonym/dialogue-QA, machine-checkable).
Contents
rl_tasks.jsonl (9583 rows)
Format
JSON Lines (.jsonl), one example per line.
Provenance
Synthetic, generated programmatically for the LemonSeed 1.5B project (by Geramy L. Loveless). Data authored by Michael Anthony Falabella.
openllm-manual-eval-codeSample conversations with vicuna-13b-fp16-v1.0 , focusing on coding ability tests. (Note that vicuna-13b have a v1.1 released and it improved on some coding tasks)
Probably not going to make more like this as creating JSON file by hand (copy-pasting from my note app) is exceedingly slow.