SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2503.05142 · ICLR · 2025

RocketEval: Efficient Automated LLM Evaluation via Grading Checklist

Xing Sun, Tianjun Wei, Wei Wen, Ruizhi Qiao, Jianghong Ma

arXiv · PDF · Open in the Atlas

Code that ran

4 of 5 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; 0 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 4 failures of our instrument; 1 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/2503.05142.json · badge: /api/ran/2503.05142.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
joinn99/rocketeval-iclr canonical (classifier label)
classifier basis: classifier_2025
4 ran · 1 no recorded run
FunctionStatusWhere it lives
chat_history Ran
ran; our contract draft was wrong, not the code
joinn99/rocketeval-iclr/src/rocketeval/process/checklist.py
code served (permissive licence) · get_code("b13720137bf9399f")
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 torch

# Construct a valid conversation_input argument
# The function expects a list of dicts with 'role' and 'content' keys
conversation_input = [
    {'role': 'user', 'content': 'Hello, how are you?'},
    {'role': 'assistant', 'content': 'I am doing well, thank you for asking.'},
    {'role': 'user', 'content': 'What is the weather today?'}
]

ARGS = {
    'conversation_input': conversation_input
}
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 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
load_bench_data Ran
ran; our contract draft was wrong, not the code
joinn99/rocketeval-iclr/src/rocketeval/process/checklist.py
code served (permissive licence) · get_code("1275ab0a44ccee37")
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 os
import pandas as pd
import tempfile
import json

# Create a temporary directory structure with mock data files
temp_dir = tempfile.mkdtemp()
data_dir = os.path.join(temp_dir, 'data')
dataset_dir = os.path.join(data_dir, 'wildbench')
checklist_dir = os.path.join(dataset_dir, 'checklist')

os.makedirs(checklist_dir, exist_ok=True)

# Create mock query.jsonl file
query_data = [
    {'session_id': '1', 'query': 'What is AI?', 'checklist': 'item1'},
    {'session_id': '2', 'query': 'How does ML work?', 'checklist': 'item2'},
]
with open(os.path.join(dataset_dir, 'query.jsonl'), 'w') as f:
    for record in query_data:
        f.write(json.dumps(record) + '\n')

# Create mock checklist.jsonl file
checklist_data = [
    {'session_id': '1', 'check1': 'pass', 'check2': 'fail'},
    {'session_id': '2', 'check1': 'pass', 'check2': 'pass'},
]
with open(os.path.join(checklist_dir, 'checklist.jsonl'), 'w') as f:
    for record in checklist_data:
        f.write(json.dumps(record) + '\n')

ARGS = {
    'data_dir': data_dir,
    'dataset_name': 'wildbench',
    'load_checklist': False,
}
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 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
prompt_to_message Ran
ran; our contract draft was wrong, not the code
joinn99/rocketeval-iclr/src/rocketeval/process/checklist.py
code served (permissive licence) · get_code("e779fa1c3f5da509")
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 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)

ARGS = {
    'prompt': 'Hello, how are you?',
    'system_message': 'You are a helpful assistant.'
}
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 list, ndim 1, 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
row_create Ran
ran; our contract draft was wrong, not the code
joinn99/rocketeval-iclr/src/rocketeval/process/checklist.py
code served (permissive licence) · get_code("a3000858ed4771a8")
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 dict. 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 torch.nn as nn

# Define ARGS dictionary with all parameters for row_create
ARGS = {
    'model': 'gpt-4',
    'custom_id': 'batch_request_001',
    'messages': [
        {'role': 'system', 'content': 'You are a helpful assistant.'},
        {'role': 'user', 'content': 'Hello, how are you?'}
    ],
    'temperature': 0.7,
    'max_tokens': 1024,
    'logprobs': False,
    'top_logprobs': 5,
    'top_p': 0.95
}
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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict 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
create_checklist Not yet run
unverified (harvested, no recorded run)
joinn99/rocketeval-iclr/src/rocketeval/process/checklist.py
code served (permissive licence) · get_code("70d43bc3b7e3389f")
How it ran

no invocation recorded.

Also present in other papers' code

Of 5 samples lifted from this paper's linked repositories, 5 carry a function identity and 0 do not. Of the 5 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 5 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

Evaluating large language models (LLMs) in diverse and challenging scenarios is essential to align them with human preferences. To mitigate the prohibitive costs associated with human evaluations, utilizing a powerful LLM as a judge has emerged as a favored approach. Nevertheless, this methodology encounters several challenges, including substantial expenses, concerns regarding privacy and security, and reproducibility. In this paper, we propose a straightforward, replicable, and accurate automated evaluation method by leveraging a lightweight LLM as the judge, named RocketEval. Initially, we identify that the performance disparity between lightweight and powerful LLMs in evaluation tasks primarily stems from their ability to conduct comprehensive analyses, which is not easily enhanced through techniques such as chain-of-thought reasoning. By reframing the evaluation task as a multi-faceted Q&A using an instance-specific checklist, we demonstrate that the limited judgment accuracy of lightweight LLMs is largely attributes to high uncertainty and positional bias. To address these challenges, we introduce an automated evaluation process grounded in checklist grading, which is designed to accommodate a variety of scenarios and questions. This process encompasses the creation of checklists, the grading of these checklists by lightweight LLMs, and the reweighting of checklist items to align with the supervised annotations. Our experiments carried out on the automated evaluation benchmarks, MT-BENCH and WILDBENCH datasets, reveal that RocketEval, when using Gemma-2-2B as the judge, achieves a high correlation (0.965) with human preferences, which is comparable to GPT-4o. Moreover, RocketEval provides a cost reduction exceeding 50-fold for large-scale evaluation and comparison scenarios. Our code is available at https://github.com/Joinn99/RocketEval-ICLR.

For agents

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

get_harvested_code_for_paper("2503.05142")
get_code_for_paper("2503.05142")
have("2503.05142")

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

curl https://syntology.ai/api/ran/2503.05142.json

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

[![Syntology run record](https://syntology.ai/api/ran/2503.05142.svg)](https://syntology.ai/paper/2503.05142)

Connect an agent — have() is free.