Base collapses on multi-digit multiplication (often copies an operand or invents digits). R5b runs full abacus-style traces (base-1000 chunked mult for ≥10-digit products) and finishes with a parseable expr=result line. Head-to-head numbers above are the R4 release suite (same seed); stratified R5b below is stronger on hard / hard_div / xl.
Stratified suite (greedy, seed=123, eval_math.py; Base from the same harness as our earlier deploy report):
Suite
Base
Zhusuan R5b
Δ
easy (n=100)
96.0%
99.0%
+3.0
medium (n=100)
90.0%
98.0%
+8.0
hard (n=80)
63.7%
100%
+36.3
hard — multiplication only
54.8% (23/42)
100% (40/40)
+45.2
hard — division only
83.3% (10/12)
100% (13/13)
+16.7
hard_div (large divisors, n=40)
—
97.5%
new stress set
xl_fixed (n=30)
83.3%
100%
+16.7
plain_easy (n=50)
—
98.0%
no abacus cue
Overall (easy+med+hard+xl, n=310)
84.5%
99.0%
+14.5
R4 vs Base — full chart
Base tends to slip digits on long multiplication, contradict itself, and emit hard-to-parse final lines — real outputs in Demo. Eval is synthetic abacus-style arithmetic (code in sft/); competition word problems are not the target. Repo weight is R5b (chunked big-mult; init R5a → R5b).
Quick start
No system prompt needed — the bundled chat template injects a zhusuan one by default. Ask in any language (用珠算思路算 73×19, Calculate 73×19 …). Use greedy decoding and max_new_tokens ≥ 3072 (long-multiplication traces are long). The answer is the last line after </think>.
LFM Open License v1.0(LICENSE 参照)。年間売上 1,000 万米ドル以上の主体による商用利用は本契約では許諾されません。
Demo / 示例 / デモ
Real greedy generations. Note: the model always thinks in its trained Chinese abacus trace, even for English/Japanese prompts.(模型无论用什么语言提问,都用中文珠算轨迹思考。どの言語で聞いても中国語の珠算トレースで考えます。)
Calculate 73×19 using abacus-style (珠算) reasoning. — full output
Calculate 35937×1774 using abacus-style (珠算) reasoning. — base vs this model
Base burns its thinking budget guessing what 珠算 means, then slips a digit → wrong:
text
1<think> Okay, let's tackle this problem ... Hmm, first, I need to recall what
2abacus-style calculation involves. Since the user mentioned using the abacus
3and pearls (珠算), which I think refers to the traditional Chinese counting
4method ... </think>
5... After precise computation, the result is **63,753,238**.
67\boxed{63753238} ← digit slip, WRONG (correct: 63752238)
(Japanese prompt, same story: base calls it the "Jewel Box Suanpan method" and answers 63,753,238.)
This model (EN or JA prompt — identical output) runs the full partial-product trace (abridged) → correct: