SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2306.05087 · ICLR · 2024

PandaLM: An Automatic Evaluation Benchmark for LLM Instruction Tuning Optimization

Jindong Wang, Hao Chen, Xing Xie, Yue Zhang, Yidong Wang, Linyi Yang, Wei Ye, Shikun Zhang, Cunxiang Wang, Zhuohao Yu, Wenjin Yao, Zhengran Zeng, and 2 more

arXiv · PDF · Open in the Atlas

Code that ran

3 of 3 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; 2 failures of our instrument; 0 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/2306.05087.json · badge: /api/ran/2306.05087.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
weopenml/pandalm canonical (classifier label)
classifier basis: classifier_2025
2 ran · 0 no recorded run
baaivision/judgelm reimplementation (classifier label)
classifier basis: classifier_2025
1 ran · 0 no recorded run
FunctionStatusWhere it lives
build_prompt Ran
ran; our contract draft was wrong, not the code
weopenml/pandalm/pandalm/run-gradio.py
code served (permissive licence) · get_code("f4bb8eecfe1a4f76")
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

ARGS = {
    'instruction': 'Summarize the following text in one sentence.',
    'input': 'The quick brown fox jumps over the lazy dog. This is a common pangram used in typography.',
    'resp1': 'A fast brown fox jumps over a sleeping dog.',
    'resp2': 'The agile fox leaps across an inactive canine.',
    'result': 'Response 1 is more concise and accurate.',
    'explain': 'Response 1 captures the essential elements of the text more clearly.',
    'ref': 'A quick brown fox jumps over a lazy dog, which is a pangram.'
}
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
filt_pred_score_list_by_gt_score_list Ran
ran, honoured the contract we drafted
baaivision/judgelm/judgelm/llm_judge/eval_model_judgement.py
code served (permissive licence) · get_code("b904d307320eb547")
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 numpy as np

torch.manual_seed(0)

# The function filters two lists based on whether gt_score contains -1
# gt_score_list: list of score tuples/lists where each has at least 2 elements
# pred_score_list: list of corresponding pred scores (can be any type)

gt_score_list = [
    [0.5, 0.7],
    [-1, 0.3],
    [0.2, 0.4],
    [0.1, -1],
    [0.9, 0.8],
]

pred_score_list = [
    0.6,
    0.4,
    0.3,
    0.2,
    0.85,
]

ARGS = {
    "gt_score_list": gt_score_list,
    "pred_score_list": pred_score_list,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, 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
post_process_output Ran
ran; our contract draft was wrong, not the code
weopenml/pandalm/pandalm/run-gradio.py
code served (permissive licence) · get_code("14bd70346084e2e2")
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 str. 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 re

def post_process_output(text):
    text = text.strip().split('### Evaluation:')[1].strip()
    pattern = re.compile('<unk>|<pad>|<s>|</s>|\\[PAD\\]|<\\|endoftext\\|>|\\[UNK\\]|\\[CLS\\]|\\[MASK\\]|<\\|startofpiece\\|>|<\\|endofpiece\\|>|\\[gMASK\\]|\\[sMASK\\]')
    pattern.sub('', text.strip()).strip()
    return text

ARGS = {
    "text": "Some prefix text\n### Evaluation:\nThis is the evaluation result with <unk> and [PAD] tokens to remove."
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 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

Also present in other papers' code

Of 3 samples lifted from this paper's linked repositories, 3 carry a function identity and 0 do not. Of the 3 distinct functions with an identity, 1 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 1 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.

FunctionAlso present inEarliest-datedOthers named
filt_pred_score_list_by_gt_score_list 3 other papers
3 later
2310.17631 (2023-10)
JudgeLM: Fine-tuned Large Language Models are Scalable Judges
2024.findings-emnlp.559 (2024 (year only))
A LLM-Based Ranking Method for the Evaluation of Automatic Counter-Narrative Generation
2409.16788 (2024-09)
Mitigating the Bias of Large Language Model Evaluation

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

Instruction tuning large language models (LLMs) remains a challenging task, owing to the complexity of hyperparameter selection and the difficulty involved in evaluating the tuned models. To determine the optimal hyperparameters, an automatic, robust, and reliable evaluation benchmark is essential. However, establishing such a benchmark is not a trivial task due to the challenges associated with evaluation accuracy and privacy protection. In response to these challenges, we introduce a judge large language model, named PandaLM, which is trained to distinguish the superior model given several LLMs. PandaLM's focus extends beyond just the objective correctness of responses, which is the main focus of traditional evaluation datasets. It addresses vital subjective factors such as relative conciseness, clarity, adherence to instructions, comprehensiveness, and formality. To ensure the reliability of PandaLM, we collect a diverse human-annotated test dataset, where all contexts are generated by humans and labels are aligned with human preferences. On evaluations using our collected test dataset, our findings reveal that PandaLM-7B offers performance comparable to both GPT-3.5 and GPT-4. Impressively, PandaLM-70B surpasses their performance. PandaLM enables the evaluation of LLM to be fairer but with less cost, evidenced by significant improvements achieved by models tuned through PandaLM compared to their counterparts trained with default Alpaca's hyperparameters. In addition, PandaLM does not depend on API-based evaluations, thus avoiding potential data leakage.

For agents

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

get_harvested_code_for_paper("2306.05087")
get_code_for_paper("2306.05087")
have("2306.05087")

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

curl https://syntology.ai/api/ran/2306.05087.json

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

[![Syntology run record](https://syntology.ai/api/ran/2306.05087.svg)](https://syntology.ai/paper/2306.05087)

Connect an agent — have() is free.