SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2607.20558 · 2026

StabilityBench: a Framework for Benchmarking Instability in Large Language Models

Reihaneh Rabbany, Zachary Yang, Anne Imouza, Emma Kondrup

arXiv · PDF

Code that ran

2 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 0 have no contract verdict; 2 failures of our instrument; 4 no recorded run), measured 2026-11-22; 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/2607.20558.json · badge: /api/ran/2607.20558.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
ekmpa/StabilityBench community
Papers with Code archive
2 ran · 4 no recorded run
FunctionStatusWhere it lives
bench_retry Ran
ran; our contract draft was wrong, not the code
ekmpa/StabilityBench/utils/bench_loaders.py
pointer only (licence: NONE) · get_code("a67f20ae116c6163")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, ekmpa/StabilityBench when it was checked. 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.

load_strongreject Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
ekmpa/StabilityBench/utils/bench_loaders.py
pointer only (licence: NONE) · get_code("9bd17f04065f9590")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, ekmpa/StabilityBench when it was checked. 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.

load_AIME Not yet run
unverified (harvested, no recorded run)
ekmpa/StabilityBench/utils/bench_loaders.py
pointer only (licence: NONE) · get_code("879357fb7ed31b39")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, ekmpa/StabilityBench when it was checked. 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. No invocation recorded.

load_benchmark Not yet run
unverified (harvested, no recorded run)
ekmpa/StabilityBench/utils/bench_loaders.py
pointer only (licence: NONE) · get_code("e997c1a44311802c")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, ekmpa/StabilityBench when it was checked. 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. No invocation recorded.

load_gsm8k Not yet run
unverified (harvested, no recorded run)
ekmpa/StabilityBench/utils/bench_loaders.py
pointer only (licence: NONE) · get_code("46502769b16b0f67")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, ekmpa/StabilityBench when it was checked. 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. No invocation recorded.

load_healthBench Not yet run
unverified (harvested, no recorded run)
ekmpa/StabilityBench/utils/bench_loaders.py
pointer only (licence: NONE) · get_code("96d0fb9e21a3289f")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, ekmpa/StabilityBench when it was checked. 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. No invocation 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, 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 6 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

AI Assistants are increasingly deployed in highstakes settings, such as healthcare or government services. Yet their real-world behavior remains poorly understood due to strong context dependence. Current evaluation protocols follow a defense-in-depth paradigm with compounding layers of safeguards, ranging from traditional benchmarks to live or adversarial testing. Such benchmarks remain largely static and single-turn, limiting their ability to capture real-world variability in conversational settings. We propose Sta-bilityBench, a principled, general and modelagnostic benchmark operator that transforms single-turn benchmark queries into multi-turn interaction histories. StabilityBench augments existing benchmarks by injecting realistic user simulations, through demographic proxies or sycophantic baits, while preserving original task intent. We apply StabilityBench to four benchmarks spanning mathematical reasoning, health questionanswering and safety, and evaluate nine large language models under these conditions. Our results show that model performance is consistently unstable under these injections, with considerable performance degradations on three out of four benchmarks studied. These highlight important limitations of static evaluations and motivate more realistic evaluation settings. To this end, we propose StabilityBench-Mini: a size-preserving variant of StabilityBench that samples across diversification axes, enabling more realistic evaluation without increasing costs.

For agents

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

get_harvested_code_for_paper("2607.20558")
get_code_for_paper("2607.20558")
have("2607.20558")

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

curl https://syntology.ai/api/ran/2607.20558.json

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

[![Syntology run record](https://syntology.ai/api/ran/2607.20558.svg)](https://syntology.ai/paper/2607.20558)

Connect an agent — have() is free.