SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2307.09009 · 2023

How is ChatGPT's behavior changing over time?

arXiv · PDF · Open in the Atlas

Code that ran

6 of 6 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 4 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/2307.09009.json · badge: /api/ran/2307.09009.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
lchen001/llmdrift canonical (classifier label)
classifier basis: classifier_2025
1 ran · 0 no recorded run
JesusCarr/JARVIS community (classifier label)
classifier basis not recorded on the link
5 ran · 0 no recorded run
FunctionStatusWhere it lives
add_text Ran
ran on a synthesized input
JesusCarr/JARVIS/server/run_gradio_demo.py
pointer only (licence: MIT) · get_code("e4b538a226989715")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, JesusCarr/JARVIS, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

configparse Ran
ran; our contract draft was wrong, not the code
lchen001/llmdrift/src/monitor/llmmonitor.py
code served (permissive licence) · get_code("59f51faca029acbd")
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 torch.nn as nn

torch.manual_seed(0)

# Build minimal config dict with all required keys
config = {
    'query_params': {
        'temperature': 0.7,
        'max_tokens': 100,
        'top_p': 0.9
    },
    'model_list': ['model_a', 'model_b', 'model_c'],
    'model_params': {
        'hidden_size': 256,
        'num_layers': 2,
        'dropout': 0.1
    },
    'trail': {
        'experiment_id': 'test_001',
        'seed': 42,
        'batch_size': 32
    },
    'system_prompts': [
        "You are a helpful assistant.",
        "You are an expert AI.",
        None
    ]
}

ARGS = {
    'config': config
}
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
extract_medias Ran
ran on a synthesized input
JesusCarr/JARVIS/server/run_gradio_demo.py
pointer only (licence: MIT) · get_code("593424d12e7135ec")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, JesusCarr/JARVIS, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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_max_context_length Ran
ran on a synthesized input
JesusCarr/JARVIS/server/get_token_ids.py
pointer only (licence: MIT) · get_code("00ead7669a85e810")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, JesusCarr/JARVIS, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

replace_slot Ran
ran; our contract draft was wrong, not the code
JesusCarr/JARVIS/server/awesome_chat.py
pointer only (licence: MIT) · get_code("d9f3c3ff8e21a46f")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, JesusCarr/JARVIS, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

set_openai_key Ran
ran on a synthesized input
JesusCarr/JARVIS/server/run_gradio_demo.py
pointer only (licence: MIT) · get_code("c198289d7525e962")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, JesusCarr/JARVIS, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

Also present in other papers' code

Of 6 samples lifted from this paper's linked repositories, 6 carry a function identity and 0 do not. Of the 6 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 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.

FunctionAlso present inEarliest-datedOthers named
replace_slot 1 other paper
1 dated earlier
2303.17580 (2023-03)
HuggingGPT: Solving AI Tasks with ChatGPT and its Friends in Hugging Face

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

GPT-3.5 and GPT-4 are the two most widely used large language model (LLM) services. However, when and how these models are updated over time is opaque. Here, we evaluate the March 2023 and June 2023 versions of GPT-3.5 and GPT-4 on several diverse tasks: 1) math problems, 2) sensitive/dangerous questions, 3) opinion surveys, 4) multi-hop knowledge-intensive questions, 5) generating code, 6) US Medical License tests, and 7) visual reasoning. We find that the performance and behavior of both GPT-3.5 and GPT-4 can vary greatly over time. For example, GPT-4 (March 2023) was reasonable at identifying prime vs. composite numbers (84% accuracy) but GPT-4 (June 2023) was poor on these same questions (51% accuracy). This is partly explained by a drop in GPT-4's amenity to follow chain-of-thought prompting. Interestingly, GPT-3.5 was much better in June than in March in this task. GPT-4 became less willing to answer sensitive questions and opinion survey questions in June than in March. GPT-4 performed better at multi-hop questions in June than in March, while GPT-3.5's performance dropped on this task. Both GPT-4 and GPT-3.5 had more formatting mistakes in code generation in June than in March. We provide evidence that GPT-4's ability to follow user instructions has decreased over time, which is one common factor behind the many behavior drifts. Overall, our findings show that the behavior of the "same" LLM service can change substantially in a relatively short amount of time, highlighting the need for continuous monitoring of LLMs.

For agents

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

get_harvested_code_for_paper("2307.09009")
get_code_for_paper("2307.09009")
have("2307.09009")

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

curl https://syntology.ai/api/ran/2307.09009.json

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

[![Syntology run record](https://syntology.ai/api/ran/2307.09009.svg)](https://syntology.ai/paper/2307.09009)

Connect an agent — have() is free.