|
Download README.md from IFM/latent-feedback-data: direct link, hf CLI and curl.
- Browser
- Download file 2 kB
-
https://huggingface.co/datasets/IFM/latent-feedback-data/resolve/main/README.md
- Command line
-
hf download hf://datasets/IFM/latent-feedback-data/README.md
-
curl -L -o README.md https://huggingface.co/datasets/IFM/latent-feedback-data/resolve/main/README.md
2 kB
| license: odc-by | |
| language: | |
| - en | |
| pretty_name: Latent Feedback Is a Compute and Data Amplifier, training data | |
| tags: | |
| - dolma | |
| - pretraining | |
| - scaling-laws | |
| # Latent Feedback Is a Compute and Data Amplifier: training data | |
| The data of *Latent Feedback Is a Compute and Data Amplifier* (paper: [alphaXiv](https://www.alphaxiv.org/abs/2611.latent-feedback-transformers-scaling-law), code: [ifm-ai/LoRM](https://github.com/ifm-ai/LoRM), folder [`works/latent_feedback`](https://github.com/ifm-ai/LoRM/tree/main/works/latent_feedback)). | |
| ## `d3/`: the training stream | |
| D3 is the [Dolma 3 mix](https://huggingface.co/datasets/allenai/dolma3_mix-150B-1025) (revision `afa92bf`), tokenized | |
| with the Llama 2 tokenizer, then processed in three steps: | |
| 1. **Deduplicated.** Exact duplicate documents are removed, keeping the first copy. The mix repeats documents, so about | |
| 60% of its tokens are extra copies. | |
| 2. **Stratified.** Documents are put in a seed-0 nested order over the mix's 224 source-topic buckets. Every prefix | |
| follows the mixture of the whole, so a smaller unique-data pool is a prefix of a larger one. | |
| 3. **Decontaminated.** Training documents that share a paragraph of 13 or more words with any Paloma source (val and | |
| test) are dropped. | |
| | Split | Shards | Tokens | Documents | | |
| | --- | --- | --- | --- | | |
| | train | 757 | 75.6B | 35,576,659 | | |
| | val | 1 | 100.0M | 47,274 | | |
| - **Shard format:** each document is BOS + tokens + EOS, as uint16 after a 1,024-byte header (`loom.data`). | |
| - **`manifest.json`:** shard sizes, bucket mixture and duplicate statistics. | |
| - **`removed.npz`:** the documents dropped by decontamination. | |
| `python -m works.latent_feedback.data` rebuilds D3 from the mix. | |
| ## `task-bpb/`: the task requests | |
| These are the requests of the Olmo 3 Base Easy BPB suites (math, code and QA; 109 tasks), built with OLMES. There is one | |
| JSONL per task with context, continuation and gold label. `measure/task_bpb.py` scores them. Each task keeps its source | |
| dataset's license. | |