Part of the LittleLearner scale-up study (pedagogically-controlled knowledge exposure): Qwen3 dense LMs trained on a corpus filtered to U.S. K-5 material (bounded) vs an unfiltered FineWeb-Edu corpus (unbounded), to measure what an interpretable knowledge boundary costs and grants.

Model

Architecture: Qwen3 dense (Qwen3ForCausalLM).
Size: 5.04B params, hidden 3072, 44 layers, 24 query / 8 KV heads, FFN 9216. Context: 4096.
Tokenizer: custom 64k byte-level BPE with per-digit splitting (ChatML special tokens).
Pretraining: 88B tokens on K-5 LittleCurriculum (FineWeb-Edu filtered to U.S. grades K-5). WSD schedule, sharded Muon, MXFP8, Megatron-Core on 8xB200.

Evaluation

In-domain bits-per-byte (BPB): 0.536.

Usage

python

# transformers (completion)
from transformers import AutoModelForCausalLM, AutoTokenizer
repo = "manueldeprada/littlelearner-5b-bounded-base"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(repo, dtype="bfloat16", device_map="cuda")
ids = tok("The sum of 2 and 3 is", return_tensors="pt").to(model.device)
print(tok.decode(model.generate(**ids)[0], skip_special_tokens=True))

python

# vLLM
from vllm import LLM
llm = LLM("manueldeprada/littlelearner-5b-bounded-base")
print(llm.generate(["The sum of 2 and 3 is"])[0].outputs[0].text)

littlelearner/littlelearner-5b-base

littlelearner-5b-base

Model

Evaluation

Usage

littlelearner/
littlelearner-5b-base