SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2601.16447 · 2026

Mixing Expert Knowledge: Bring Human Thoughts Back To the Game of Go

Dahua Lin, Kai Chen, Qipeng Guo, Jiasheng Ye, Linyang Li, Yongkang Chen, Yichuan Ma, Peiji Li

arXiv · PDF · Open in the Atlas

Code that ran

2 of 2 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 1 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 1 failure of our instrument; 0 no recorded run), measured 2026-11-22; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2601.16447.json · badge: /api/ran/2601.16447.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
Entarochuan/LoGos community
Papers with Code archive
2 ran · 0 no recorded run
FunctionStatusWhere it lives
compute_score Ran
ran, honoured the contract we drafted
Entarochuan/LoGos/RL_utils/Go_reward.py
code served (permissive licence) · get_code("6026905b22ef8498")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned float, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import re
import random

# Set seed for reproducibility
random.seed(0)

# Create a valid solution_str that matches the expected format
solution_str = """<|im_start|>assistant
<reasoning>
Let me analyze this Go position carefully.
The board state suggests we should play at position E10.
</reasoning>

<answer>
\\boxed{下一步颜色:黑}
\\boxed{下一步位置:E10}
\\boxed{下一步胜率:65.5\\%}
</answer>
"""

# Create a ground_truth dict with the expected structure
ground_truth = {
    'former_moves': ['D4', 'D16', 'P16'],  # 3 moves, so next move is move 4 (white's turn)
    'candidates': [
        {
            'move': 'E10',
            'win_rate': 0.655
        },
        {
            'move': 'F11',
            'win_rate': 0.640
        },
        {
            'move': 'G12',
            'win_rate': 0.620
        },
        {
            'move': 'H13',
            'win_rate': 0.600
        },
        {
            'move': 'I14',
            'win_rate': 0.580
        },
        {
            'move': 'J15',
            'win_rate': 0.560
        },
        {
            'move': 'K16',
            'win_rate': 0.540
        },
        {
            'move': 'L17',
            'win_rate': 0.520
        },
        {
            'move': 'M18',
            'win_rate': 0.500
        },
        {
            'move': 'N19',
            'win_rate': 0.480
        }
    ]
}

ARGS = {
    'solution_str': solution_str,
    'ground_truth': ground_truth
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
extract_coordinate Ran
ran; our contract draft was wrong, not the code
Entarochuan/LoGos/RL_utils/Go_reward.py
code served (permissive licence) · get_code("c1448fec3dcf7453")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1) and returned tuple, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import re

ASSISTANT_PATTERN = re.compile('<\\|im_start\\|>assistant\\n(.*)', re.DOTALL)
REASONING_PATTERN = re.compile('<reasoning>(.*?)</reasoning>', re.DOTALL)

def extract_coordinate(solution_str):
    """
    从模型输出中提取坐标和颜色
    处理步骤:
    1. 提取assistant回复部分
    2. 检查是否存在<think>标签
    3. 从<answer>标签或原始格式中提取坐标和颜色
    """
    try:
        assistant_match = ASSISTANT_PATTERN.search(solution_str)
        if not assistant_match:
            content = solution_str
        else:
            content = assistant_match.group(1)
        has_think = bool(REASONING_PATTERN.search(content))
        reasoning_match = REASONING_PATTERN.search(content)
        if reasoning_match:
            reasoning_content = reasoning_match.group(1)
        else:
            reasoning_content = ''
        answer_match = re.search('<answer>(.*?)</answer>', content, re.DOTALL)
        if answer_match:
            content = answer_match.group(1)
        else:
            return None
        color_match = re.search('\\\\boxed\\{下一步颜色:(黑|白)\\}', content)
        if not color_match:
            return None
        color = color_match.group(1)
        coord_match = re.search('\\\\boxed\\{下一步位置:([A-HJ-T]\\d+)\\}', content)
        if not coord_match:
            coord_match = re.search('\\\\boxed\\{([A-HJ-T]\\d+)\\}', content)
            if not coord_match:
                return None
        win_rate_match = re.search('\\\\boxed\\{下一步胜率:(\\d+\\.\\d+)\\%\\}', content)
        if not win_rate_match:
            return None
        win_rate = float(win_rate_match.group(1))
        coordinate = coord_match.group(1)
        letter, number = (coordinate[0], int(coordinate[1:]))
        if 1 <= number <= 19:
            return (coordinate, color, win_rate, has_think, reasoning_content)
        return None
    except Exception as e:
        print(f'Error extracting coordinate and color: {str(e)}')
        return None

# Valid test case with all required patterns
solution_str = (
    '<|im_start|>assistant\n'
    '<reasoning>Let me think about this move</reasoning>\n'
    '<answer>\n'
    '\\boxed{下一步颜色:黑}\n'
    '\\boxed{下一步位置:A1}\n'
    '\\boxed{下一步胜率:75.50%}\n'
    '</answer>'
)

ARGS = {
    'solution_str': solution_str
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family float — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block

Also present in other papers' code

Of 2 samples lifted from this paper's linked repositories, 2 carry a function identity and 0 do not. Of the 2 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 2 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Large language models (LLMs) have demonstrated exceptional performance in reasoning tasks such as mathematics and coding, matching or surpassing human capabilities. However, these impressive reasoning abilities face significant challenges in specialized domains. Taking Go as an example, although AlphaGo has established the high performance ceiling of AI systems in Go, mainstream LLMs still struggle to reach even beginner-level proficiency, let alone perform natural language reasoning. This performance gap between general-purpose LLMs and domain experts is significantly limiting the application of LLMs on a wider range of domain-specific tasks. In this work, we aim to bridge the divide between LLMs' general reasoning capabilities and expert knowledge in domain-specific tasks. We perform mixed fine-tuning with structured Go expertise and general long Chain-of-Thought (CoT) reasoning data as a cold start, followed by reinforcement learning to integrate expert knowledge in Go with general reasoning capabilities. Through this methodology, we present LoGos, a powerful LLM that not only maintains outstanding general reasoning abilities, but also conducts Go gameplay in natural language, demonstrating effective strategic reasoning and accurate next-move prediction. Lo-Gos achieves performance comparable to human professional players, substantially surpassing all existing LLMs. Through this work, we aim to contribute insights on applying general LLM reasoning capabilities to specialized domains. We will release the first large-scale Go dataset for LLM training, the first LLM Go evaluation benchmark, and the first general LLM that reaches human professional-level performance in Go at: https://github.com/Entarochuan/LoGos.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2601.16447")
get_code_for_paper("2601.16447")
have("2601.16447")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2601.16447.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2601.16447.svg)](https://syntology.ai/paper/2601.16447)

Connect an agent — have() is free.