93 lines
3.8 KiB
Markdown
93 lines
3.8 KiB
Markdown
---
|
|
library_name: transformers
|
|
tags:
|
|
- text-generation
|
|
- legal
|
|
- finance
|
|
- pretraining
|
|
- from-scratch
|
|
pipeline_tag: text-generation
|
|
---
|
|
|
|
# SLM125 — a 125M-parameter legal/financial base language model, trained from scratch
|
|
|
|
**SLM125** is a small (~125.8M parameter), decoder-only transformer pretrained from
|
|
scratch on a legal- and finance-heavy corpus. It is a **base** model — it completes
|
|
text, it is not instruction-tuned or chat-tuned. Give it a sentence to continue
|
|
("The plaintiff argued that ..."), not a question.
|
|
|
|
🔗 **Try it live:** [SLM125 Playground](https://web-roan-chi-97.vercel.app)
|
|
|
|
## Model Details
|
|
|
|
- **Architecture:** Llama-style decoder-only transformer (RoPE, RMSNorm, SwiGLU MLP, tied embeddings)
|
|
- **Parameters:** ~125.8M
|
|
- **Hidden size:** 768
|
|
- **Layers:** 12
|
|
- **Attention heads:** 12 (head dim 64, standard multi-head attention — no GQA)
|
|
- **Intermediate (MLP) size:** 3,072
|
|
- **Context length:** 1,024 tokens
|
|
- **Vocabulary:** 16,384 tokens (custom-trained tokenizer)
|
|
- **RoPE theta:** 10,000
|
|
- **Precision:** trained and served in fp32/bf16 on Modal
|
|
|
|
## Training Data
|
|
|
|
Pretrained on a cleaned, deduplicated, decontaminated corpus of **~2.19B tokens**,
|
|
mixed "legal-first" from three public, streamed HuggingFace datasets:
|
|
|
|
| Source | Dataset | Share | What it is |
|
|
|---|---|---|---|
|
|
| case-law | [`HFforLegal/case-law`](https://huggingface.co/datasets/HFforLegal/case-law) (`us` config) | ~39% (~863M tokens) | US court opinions (scanned; some OCR noise) |
|
|
| sec | [`PleIAs/SEC`](https://huggingface.co/datasets/PleIAs/SEC) | ~39% (~861M tokens) | SEC filings (10-K, etc.), born-digital |
|
|
| fineweb-edu | [`HuggingFaceFW/fineweb-edu`](https://huggingface.co/datasets/HuggingFaceFW/fineweb-edu) (`sample-10BT`) | ~21% (~465M tokens) | General educational web text, added as fluency filler |
|
|
|
|
The two legal sources were taken in full (they cap out around 2B tokens combined);
|
|
a smaller web slice was added on top, landing at roughly a 40/40/20 split — about
|
|
78% legal/financial text overall, not the originally-planned 70/20/10.
|
|
|
|
## How to Get Started
|
|
|
|
```python
|
|
from transformers import AutoModelForCausalLM, AutoTokenizer
|
|
|
|
repo = "srinivasch87/slm125live-base"
|
|
tokenizer = AutoTokenizer.from_pretrained(repo)
|
|
model = AutoModelForCausalLM.from_pretrained(repo)
|
|
|
|
prompt = "The plaintiff argued that"
|
|
inputs = tokenizer(prompt, return_tensors="pt")
|
|
output = model.generate(
|
|
**inputs,
|
|
max_new_tokens=128,
|
|
do_sample=True,
|
|
temperature=0.8,
|
|
top_p=0.95,
|
|
top_k=50,
|
|
)
|
|
print(tokenizer.decode(output[0][inputs["input_ids"].shape[1]:], skip_special_tokens=True))
|
|
```
|
|
|
|
## Intended Use & Limitations
|
|
|
|
- This is a **base completion model**, not an instruction-following assistant.
|
|
It will not reliably answer questions, follow chat-style instructions, or
|
|
refuse unsafe requests — prompt it with the start of a sentence or document.
|
|
- At 125M parameters and ~2.2B training tokens, factual accuracy, coherence over
|
|
long spans, and reasoning are limited. Treat all output as unverified
|
|
autocomplete, not legal, financial, or professional advice.
|
|
- Training data skews heavily toward US case law and SEC filings, so the model's
|
|
style and any biases reflect that domain (and the OCR noise present in the
|
|
scanned case-law source).
|
|
- License is not specified for this model; the underlying training datasets each
|
|
carry their own terms — check the dataset cards linked above before
|
|
redistributing outputs.
|
|
|
|
## Training & Serving Infrastructure
|
|
|
|
Built end-to-end on [Modal](https://modal.com): data cleaning, deduplication,
|
|
tokenizer training, and pretraining all ran as fanned-out CPU/GPU Modal functions
|
|
against a shared Modal Volume. Inference is served from a Modal web endpoint with
|
|
token-streaming generation, behind a Next.js frontend deployed on Vercel (link
|
|
above).
|