SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2310.01377 · ICML · 2024

ULTRAFEEDBACK: Boosting Language Models with Scaled AI Feedback

Maosong Sun, Zhiyuan Liu, Ning Ding, Yankai Lin, Wei Zhu, Ganqu Cui, Lifan Yuan, Bingxiang He, Guanming Yao, Yuan Ni, Guotong Xie, Ruobing Xie, and 1 more

arXiv · PDF · Open in the Atlas

Code that ran

5 of 7 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; 2 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 3 failures of our instrument; 2 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/2310.01377.json · badge: /api/ran/2310.01377.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
thunlp/ultrafeedback canonical (classifier label)
classifier basis: census
5 ran · 2 no recorded run
FunctionStatusWhere it lives
calculate_average_rating Ran
ran, honoured the contract we drafted
thunlp/ultrafeedback/src/data_annotation/fix_overall_score_issue.py
code served (permissive licence) · get_code("7d3323a5d2d6ed3e")
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 torch

ARGS = {
    'annotations': {
        'item_1': {'Rating': '5'},
        'item_2': {'Rating': '4'},
        'item_3': {'Rating': '3'},
        'item_4': {'Rating': 'N/A'},
        'item_5': {'Rating': '2'},
        'item_6': {}
    }
}
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 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
check_score Ran
ran, honoured the contract we drafted
thunlp/ultrafeedback/src/data_annotation/fix_overall_score_issue.py
code served (permissive licence) · get_code("7aabc2143f0f3e10")
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 int, ndim 0, dtype family int, 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 torch

torch.manual_seed(0)

# The function checks a 'fine-grained_score' key in a completion dict
# and returns different values based on score thresholds
completion = {
    'fine-grained_score': 3
}

ARGS = {
    'completion': completion
}
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 int, ndim 0, dtype family int, 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
get_eval Ran
ran; our contract draft was wrong, not the code
thunlp/ultrafeedback/src/data_annotation/annotate_critique.py
code served (permissive licence) · get_code("a506738814a38c14")
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 openai

# Mock the OpenAI API to avoid actual network calls
class MockOpenAIResponse:
    def __getitem__(self, key):
        if key == 'choices':
            return [{'message': {'content': 'Mock evaluation result'}}]
        return None

class MockMessage:
    def __init__(self):
        self.content = 'Mock evaluation result'

class MockChoice:
    def __init__(self):
        self.message = MockMessage()

class MockChatCompletion:
    @staticmethod
    def create(**kwargs):
        return {
            'choices': [
                {'message': {'content': 'Mock evaluation result'}}
            ]
        }

# Patch openai module
openai.ChatCompletion = MockChatCompletion()

# Define arguments for get_eval function
ARGS = {
    'model': 'gpt-4',
    'sys_prompt': 'You are a helpful assistant.',
    'user_prompt': 'What is 2 + 2?'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
openai — 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
get_eval Ran
ran; our contract draft was wrong, not the code
thunlp/ultrafeedback/src/data_annotation/fix_overall_score_issue.py
code served (permissive licence) · get_code("9f8a0e28964db553")
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 openai

# Mock the openai module to avoid actual API calls
class MockChatCompletion:
    @staticmethod
    def create(**kwargs):
        return {
            'choices': [
                {
                    'message': {
                        'content': 'A'
                    }
                }
            ]
        }

openai.ChatCompletion = MockChatCompletion()

ARGS = {
    'model': 'gpt-3.5-turbo',
    'sys_prompt': 'You are a helpful assistant.',
    'user_prompt': 'What is 2+2?'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
openai — 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
process Ran
ran; our contract draft was wrong, not the code
thunlp/ultrafeedback/src/data_annotation/annotate_preference.py
code served (permissive licence) · get_code("993cf7ccb3239873")
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 re

ARGS = {
    'responses': """Rating: 8
Rationale: The response follows the instruction well and provides accurate information.

Rating: 7
Rationale: Generally good adherence to the instruction with minor issues.

Rating: 9
Rationale: Excellent instruction following with comprehensive details.

Rating: 6
Rationale: Adequate response but could be more detailed.""",
    'aspect': 'instruction_following'
}
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 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
annotate Not yet run
unverified (harvested, no recorded run)
thunlp/ultrafeedback/src/data_annotation/annotate_critique.py
code served (permissive licence) · get_code("2b3cb1aee424f3c8")
How it ran

no invocation recorded.

get_eval Not yet run
unverified (harvested, no recorded run)
thunlp/ultrafeedback/src/data_annotation/annotate_preference.py
code served (permissive licence) · get_code("38d99fe231928f20")
How it ran

no invocation recorded.

Also present in other papers' code

Of 7 samples lifted from this paper's linked repositories, 7 carry a function identity and 0 do not. Of the 7 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 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.

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

Learning from human feedback has become a pivot technique in aligning large language models (LLMs) with human preferences. However, acquiring vast and premium human feedback is bottlenecked by time, labor, and human capability, resulting in small sizes or limited topics of current datasets. This further hinders feedback learning as well as alignment research within the open-source community. To address this issue, we explore how to go beyond human feedback and collect high-quality AI feedback automatically for a scalable alternative. Specifically, we identify scale and diversity as the key factors for feedback data to take effect. Accordingly, we first broaden instructions and responses in both amount and breadth to encompass a wider range of user-assistant interactions. Then, we meticulously apply a series of techniques to mitigate annotation biases for more reliable AI feedback. We finally present ULTRAFEEDBACK, a large-scale, high-quality, and diversified AI feedback dataset, which contains over 1 million GPT-4 feedback for 250k user-assistant conversations from various aspects. Built upon ULTRAFEEDBACK, we align a LLaMA-based model by best-of-n sampling and reinforcement learning, demonstrating its exceptional performance on chat benchmarks. Our work validates the effectiveness of scaled AI feedback data in constructing strong open-source chat language models, serving as a solid foundation for future feedback learning research.

For agents

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

get_harvested_code_for_paper("2310.01377")
get_code_for_paper("2310.01377")
have("2310.01377")

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

curl https://syntology.ai/api/ran/2310.01377.json

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

[![Syntology run record](https://syntology.ai/api/ran/2310.01377.svg)](https://syntology.ai/paper/2310.01377)

Connect an agent — have() is free.