SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2304.08247 · 2023

MedAlpaca -- An Open-Source Collection of Medical Conversational AI Models and Training Data

arXiv · PDF · Open in the Atlas

Code that ran

10 of 13 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 4 have no contract verdict; 5 failures of our instrument; 3 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; 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/2304.08247.json · badge: /api/ran/2304.08247.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
tuneinsight/federated-llms community (classifier label)
classifier basis not recorded on the link
10 ran · 3 no recorded run
FunctionStatusWhere it lives
compute_bleu_score Ran
ran on a synthesized input
tuneinsight/federated-llms/src/privacy_benchmark.py
code served (permissive licence) · get_code("0d42872657378ac1")
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, through the batch API) and returned DataFrame. The image it ran in was not recorded. If your copy returns the same type 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 pandas as pd
from nltk.translate.bleu_score import sentence_bleu

# Create a small dataframe with the required columns for compute_bleu_score
# The function expects: answer50, generation50, answer_all, generation_all columns

df = pd.DataFrame({
    'answer50': [
        ['the', 'cat', 'sat'],
        ['a', 'dog', 'ran']
    ],
    'generation50': [
        ['the', 'cat', 'sat', 'down'],
        ['a', 'dog', 'jumped']
    ],
    'answer_all': [
        ['the', 'quick', 'brown', 'cat', 'sat', 'on', 'the', 'mat'],
        ['a', 'lazy', 'dog', 'ran', 'fast']
    ],
    'generation_all': [
        ['the', 'quick', 'brown', 'cat', 'sat'],
        ['a', 'lazy', 'dog', 'ran']
    ]
})

ARGS = {
    'df': df
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
nltk, pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DataFrame — a DataFrame output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DataFrame output the recorded ndim 2 and dtype family float describe that coercion, not the output, and are not stated
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
compute_sigma Ran
ran on a synthesized input
tuneinsight/federated-llms/src/dp.py
code served (permissive licence) · get_code("1538a6ee2d50a0f1")
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, through the batch API) 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 math

ARGS = {
    "eps": 0.5,
    "delta": 1e-5,
    "c2": 1.1,
    "batch_size": 32,
    "total_samples": 1000,
    "num_iterations": 100,
    "rho": 1.0,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
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
count_PHI_memorization Ran
ran on a synthesized input
tuneinsight/federated-llms/src/privacy_benchmark.py
code served (permissive licence) · get_code("3d7b3a79249ce95c")
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, through the batch API) and returned tuple. The image it ran in was not recorded. If your copy returns the same type 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 pandas as pd
import re

# Create a small DataFrame with the required columns
df = pd.DataFrame({
    'id': ['id1', 'id2'],
    'prompt_length': [100, 150],
    'answer50': ['John Smith works at the hospital', 'Jane Doe lives in NYC'],
    'generation50': ['John Smith is a doctor', 'Jane Doe is from New York']
})

# Create metadata with PHI tags
metadata = [
    {
        'id': 'id1',
        'tags': [
            {'value': 'John Smith', 'type': 'PERSON'},
            {'value': 'hospital', 'type': 'LOCATION'}
        ]
    },
    {
        'id': 'id2',
        'tags': [
            {'value': 'Jane Doe', 'type': 'PERSON'},
            {'value': 'NYC', 'type': 'LOCATION'}
        ]
    }
]

ARGS = {
    'df': df,
    'metadata': metadata
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is stated
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
count_exact_prefix_match Ran
ran on a synthesized input
tuneinsight/federated-llms/src/privacy_benchmark.py
code served (permissive licence) · get_code("2371ab764f99f368")
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, through the batch API) and returned DataFrame. The image it ran in was not recorded. If your copy returns the same type 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 pandas as pd
import torch
from transformers import AutoTokenizer

# Create a simple mock tokenizer that returns input_ids
class MockTokenizer:
    def __call__(self, texts, add_special_tokens=False):
        # Return a dict with 'input_ids' key containing token sequences
        # Simulate tokenization by converting characters to simple token ids
        if isinstance(texts, str):
            texts = [texts]
        
        input_ids = []
        for text in texts:
            # Simple mock: convert each character to a token id
            tokens = [ord(c) % 100 for c in text[:10]]  # Limit to 10 tokens
            input_ids.append(tokens)
        
        return {'input_ids': input_ids}

# Create sample dataframe with required columns
df = pd.DataFrame({
    'answer50': ['The quick brown fox', 'Hello world test'],
    'generation50': ['The quick brown cat', 'Hello world test'],
    'answer_all': ['The quick brown fox jumps over the lazy dog', 'Hello world this is a test'],
    'generation_all': ['The quick brown fox jumps over the lazy dog', 'Hello world this is a test'],
})

# Create tokenizer
tokenizer = MockTokenizer()

# Define arguments
ARGS = {
    'df': df,
    'tokenizer': tokenizer,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pandas, torch, transformers — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DataFrame — a DataFrame output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DataFrame output the recorded ndim 2 and dtype family float describe that coercion, not the output, and are not stated
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
get_auxiliary_dict Ran
ran; our contract draft was wrong, not the code
tuneinsight/federated-llms/OpenFedLLM/federated_learning/fed_utils.py
code served (permissive licence) · get_code("962adc31294b40bf")
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, through the batch API) and returned tuple. The image it ran in was not recorded. If your copy returns the same type 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 torch
import copy

torch.manual_seed(0)

class FedArgs:
    def __init__(self, fed_alg, num_clients):
        self.fed_alg = fed_alg
        self.num_clients = num_clients

# Test case 1: scaffold algorithm
fed_args_scaffold = FedArgs(fed_alg='scaffold', num_clients=3)
global_dict_scaffold = {
    'weight1': torch.randn(4, 3),
    'weight2': torch.randn(2, 5),
    'bias': torch.randn(4)
}

# Test case 2: non-scaffold algorithm
fed_args_non_scaffold = FedArgs(fed_alg='fedavg', num_clients=2)
global_dict_non_scaffold = {
    'param1': torch.randn(3, 4),
    'param2': torch.randn(5)
}

ARGS = {
    'fed_args': fed_args_scaffold,
    'global_dict': global_dict_scaffold
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is stated
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
get_clients_this_round Ran
ran, honoured the contract we drafted
tuneinsight/federated-llms/OpenFedLLM/federated_learning/fed_global.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from rui-ye/fedllm-bench, is pointer only (licence: NOASSERTION) · get_code("2128b33c10829870")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from rui-ye/fedllm-bench, and GitHub could not classify the licence of that repository (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

get_dataset_this_round Ran
ran; our contract draft was wrong, not the code
tuneinsight/federated-llms/OpenFedLLM/federated_learning/split_dataset.py
code served (permissive licence) · get_code("9c102ddaa014f4d2")
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, through the batch API) and returned Dataset. The image it ran in was not recorded. If your copy returns the same type 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 random
import torch
from datasets import Dataset

torch.manual_seed(0)
random.seed(0)

# Create a small mock dataset
# The function calls: len(dataset), dataset.select(random_idx)
# So we need a Dataset-like object with these methods
data = {
    "text": ["sample " + str(i) for i in range(20)],
    "label": [i % 2 for i in range(20)]
}
dataset = Dataset.from_dict(data)

# Round parameter
round = 0

# fed_args - not used in the function body, so minimal
class FedArgs:
    pass

fed_args = FedArgs()

# script_args - used for batch_size, gradient_accumulation_steps, max_steps
class ScriptArgs:
    batch_size = 2
    gradient_accumulation_steps = 3
    max_steps = 2

script_args = ScriptArgs()

ARGS = {
    "dataset": dataset,
    "round": round,
    "fed_args": fed_args,
    "script_args": script_args
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
datasets, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Dataset — a Dataset output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Dataset output the recorded ndim 1 and dtype family float describe that coercion, not the output, and are not stated
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
get_proxy_dict Ran
ran; our contract draft was wrong, not the code
tuneinsight/federated-llms/OpenFedLLM/federated_learning/fed_utils.py
code served (permissive licence) · get_code("38486baf4c608ec4")
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, through the batch API) 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 torch

torch.manual_seed(0)

# Create a minimal fed_args object with required attributes
class FedArgs:
    def __init__(self, fed_alg, fedopt_tau=0.1):
        self.fed_alg = fed_alg
        self.fedopt_tau = fedopt_tau

# Create a global_dict with some sample tensors
global_dict = {
    'layer1.weight': torch.randn(4, 3),
    'layer1.bias': torch.randn(4),
    'layer2.weight': torch.randn(5, 4),
}

# Test case 1: fedadagrad algorithm
fed_args_adaptive = FedArgs(fed_alg='fedadagrad', fedopt_tau=0.1)

# Test case 2: fedavgm algorithm
fed_args_avgm = FedArgs(fed_alg='fedavgm', fedopt_tau=0.1)

ARGS = {
    'fed_args': fed_args_adaptive,
    'global_dict': global_dict,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — 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
global_aggregate Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
tuneinsight/federated-llms/OpenFedLLM/federated_learning/fed_global.py
code served (permissive licence) · get_code("8fef5b18a4b5d83b")
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, through the batch API) 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 torch

torch.manual_seed(0)

# Create a simple config object with federated learning arguments
class FedArgs:
    def __init__(self, fed_alg='fedavg', num_clients=3, fedopt_beta1=0.9, fedopt_beta2=0.99, fedopt_eta=0.01, fedopt_tau=1e-8):
        self.fed_alg = fed_alg
        self.num_clients = num_clients
        self.fedopt_beta1 = fedopt_beta1
        self.fedopt_beta2 = fedopt_beta2
        self.fedopt_eta = fedopt_eta
        self.fedopt_tau = fedopt_tau

# Test with 'fedavg' algorithm
fed_args = FedArgs(fed_alg='fedavg', num_clients=3)

# Create global_dict with model parameters (tiny tensors)
global_dict = {
    'weight': torch.randn(4, 3),
    'bias': torch.randn(4)
}

# Create local_dict_list for 3 clients
local_dict_list = [
    {
        'weight': torch.randn(4, 3),
        'bias': torch.randn(4)
    },
    {
        'weight': torch.randn(4, 3),
        'bias': torch.randn(4)
    },
    {
        'weight': torch.randn(4, 3),
        'bias': torch.randn(4)
    }
]

# Sample counts per client
sample_num_list = {0: 10, 1: 15, 2: 20}

# Clients participating this round
clients_this_round = [0, 1, 2]

# Round index
round_idx = 1

# Proxy dict (for momentum-based algorithms)
proxy_dict = {
    'weight': torch.zeros(4, 3),
    'bias': torch.zeros(4)
}

# Optimizer proxy dict (for adaptive algorithms)
opt_proxy_dict = {
    'weight': torch.zeros(4, 3),
    'bias': torch.zeros(4)
}

# Auxiliary info for SCAFFOLD
auxiliary_info = (
    {'c_weight': torch.zeros(4, 3), 'c_bias': torch.zeros(4)},
    {
        0: {'c_weight': torch.randn(4, 3), 'c_bias': torch.randn(4)},
        1: {'c_weight': torch.randn(4, 3), 'c_bias': torch.randn(4)},
        2: {'c_weight': torch.randn(4, 3), 'c_bias': torch.randn(4)}
    }
)

ARGS = {
    'fed_args': fed_args,
    'global_dict': global_dict,
    'local_dict_list': local_dict_list,
    'sample_num_list': sample_num_list,
    'clients_this_round': clients_this_round,
    'round_idx': round_idx,
    'proxy_dict': proxy_dict,
    'opt_proxy_dict': opt_proxy_dict,
    'auxiliary_info': auxiliary_info
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — 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
split_dataset Ran
ran; our contract draft was wrong, not the code
tuneinsight/federated-llms/OpenFedLLM/federated_learning/split_dataset.py
code served (permissive licence) · get_code("60c6c78eb9667d17")
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, through the batch API) and returned list, ndim 2, 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 torch
import numpy as np
from datasets import Dataset

torch.manual_seed(0)

# Create a minimal dataset using HuggingFace datasets library
data = {
    'text': ['example ' + str(i) for i in range(20)],
    'label': [i % 2 for i in range(20)]
}
dataset = Dataset.from_dict(data)

# Create fed_args config object
class FedArgs:
    def __init__(self):
        self.split_strategy = 'iid'
        self.num_clients = 4

fed_args = FedArgs()

# Create script_args config object
class ScriptArgs:
    def __init__(self):
        self.seed = 42

script_args = ScriptArgs()

ARGS = {
    'fed_args': fed_args,
    'script_args': script_args,
    'dataset': dataset
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
datasets, numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 2, dtype family float — a list 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
create_PHI_injection_dataset Not yet run
unverified (harvested, no recorded run)
tuneinsight/federated-llms/src/create_dataset.py
code served (permissive licence) · get_code("e77bc1b30afa4f86")
How it ran

no invocation recorded.

create_flashcard_dataset Not yet run
unverified (harvested, no recorded run)
tuneinsight/federated-llms/src/create_dataset.py
code served (permissive licence) · get_code("940744e0891a1f3c")
How it ran

no invocation recorded.

duplicate_PHI_documents Not yet run
unverified (harvested, no recorded run)
tuneinsight/federated-llms/src/create_dataset.py
code served (permissive licence) · get_code("824e25afa42c807a")
How it ran

no invocation recorded.

Also present in other papers' code

Of 13 samples lifted from this paper's linked repositories, 13 carry a function identity and 0 do not. Of the 13 distinct functions with an identity, 6 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 6 later) and 7 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.

FunctionAlso present inEarliest-datedOthers named
get_auxiliary_dict 1 other paper
1 later
2406.04845 (2024-06)
FedLLM-Bench: Realistic Benchmarks for Federated Learning of Large Language Models
get_clients_this_round 1 other paper
1 later
2406.04845 (2024-06)
FedLLM-Bench: Realistic Benchmarks for Federated Learning of Large Language Models
get_dataset_this_round 1 other paper
1 later
2406.04845 (2024-06)
FedLLM-Bench: Realistic Benchmarks for Federated Learning of Large Language Models
get_proxy_dict 1 other paper
1 later
2406.04845 (2024-06)
FedLLM-Bench: Realistic Benchmarks for Federated Learning of Large Language Models
global_aggregate 1 other paper
1 later
2406.04845 (2024-06)
FedLLM-Bench: Realistic Benchmarks for Federated Learning of Large Language Models
split_dataset 1 other paper
1 later
2406.04845 (2024-06)
FedLLM-Bench: Realistic Benchmarks for Federated Learning of Large Language Models

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

As large language models (LLMs) like OpenAI's GPT series continue to make strides, we witness the emergence of artificial intelligence applications in an ever-expanding range of fields. In medicine, these LLMs hold considerable promise for improving medical workflows, diagnostics, patient care, and education. Yet, there is an urgent need for open-source models that can be deployed on-premises to safeguard patient privacy. In our work, we present an innovative dataset consisting of over 160,000 entries, specifically crafted to fine-tune LLMs for effective medical applications. We investigate the impact of fine-tuning these datasets on publicly accessible pre-trained LLMs, and subsequently, we juxtapose the performance of pre-trained-only models against the fine-tuned models concerning the examinations that future medical doctors must pass to achieve certification.

For agents

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

get_harvested_code_for_paper("2304.08247")
get_code_for_paper("2304.08247")
have("2304.08247")

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

curl https://syntology.ai/api/ran/2304.08247.json

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

[![Syntology run record](https://syntology.ai/api/ran/2304.08247.svg)](https://syntology.ai/paper/2304.08247)

Connect an agent — have() is free.