SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2012.01557 · ICML · 2021

Value Alignment Verification

Anca Dragan, Daniel Brown, Scott Niekum, Jordan Schneider

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; 0 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 3 failures of our instrument; 0 no recorded run), measured 2026-11-21; 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/2012.01557.json · badge: /api/ran/2012.01557.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
dsbrown1331/vav-icml canonical (classifier label)
classifier basis: classifier_2025
3 ran · 0 no recorded run
FunctionStatusWhere it lives
generate_psi Ran
ran; our contract draft was wrong, not the code
dsbrown1331/vav-icml/drivingsim_vav/algos.py
code served (permissive licence) · get_code("8c46d08023023f82")
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 ndarray, ndim 2, dtype family float, shape (2, 3) 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 numpy as np

class SimulationObject:
    def __init__(self, feed_size, num_of_features):
        self.feed_size = feed_size
        self.num_of_features = num_of_features
        self._last_input = None
    
    def feed(self, input_list):
        self._last_input = np.array(input_list)
    
    def get_features(self):
        if self._last_input is None:
            return np.zeros(self.num_of_features)
        return np.random.randn(self.num_of_features)

feed_size = 4
num_features = 3

simulation_object = SimulationObject(feed_size=feed_size, num_of_features=num_features)

inputs_set = np.random.randn(2, feed_size * 2)

ARGS = {
    "simulation_object": simulation_object,
    "inputs_set": inputs_set
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family float, shape (2, 3) — shape (2, 3) 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
information_objective_psi Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
dsbrown1331/vav-icml/drivingsim_vav/algos.py
code served (permissive licence) · get_code("7211ac6cf9bb9e60")
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 ndarray, ndim 1, dtype family float, shape (3,) 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 numpy as np

# Analyze the function to understand parameter requirements:
# - psi_set: used in w_samples.dot(psi_set.T), so needs shape (N, D) where D matches w_samples columns
# - w_samples: used in w_samples.dot(psi_set.T), shape (M, D)
# - delta_samples: reshaped to (-1, 1), used in arithmetic and sum operations, 1D array

# Create small, distinct-sized arrays
M = 2  # number of samples
D = 4  # feature dimension
N = 3  # number of psi candidates

# psi_set: shape (N, D) - candidate parameters
psi_set = np.random.randn(N, D).astype(np.float64)

# w_samples: shape (M, D) - weight samples
w_samples = np.random.randn(M, D).astype(np.float64)

# delta_samples: 1D array of length K
delta_samples = np.array([0.5, -0.3], dtype=np.float64)

ARGS = {
    "psi_set": psi_set,
    "w_samples": w_samples,
    "delta_samples": delta_samples,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family float, shape (3,) — shape (3,) 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
volume_objective_psi Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
dsbrown1331/vav-icml/drivingsim_vav/algos.py
code served (permissive licence) · get_code("c63e702a58807835")
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 ndarray, ndim 1, dtype family float, shape (3,) 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 numpy as np

# Analyze the function:
# - psi_set: used with .T (transpose), then dot product with w_samples
#   Shape should be (n_psi, d) where d matches w_samples second dimension
# - w_samples: used with .dot(psi_set.T), so shape (n_samples, d)
# - delta_samples: reshaped to (-1, 1), used in arithmetic operations
#   Shape should be (n_samples,)

# All operations are element-wise and use .sum(axis=0), suggesting
# psi_set has multiple columns (multiple psi vectors)

# Small concrete sizes:
n_samples = 2
n_psi = 3
d = 4

psi_set = np.random.randn(n_psi, d).astype(np.float64)
w_samples = np.random.randn(n_samples, d).astype(np.float64)
delta_samples = np.random.randn(n_samples).astype(np.float64)

ARGS = {
    'psi_set': psi_set,
    'w_samples': w_samples,
    'delta_samples': delta_samples,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family float, shape (3,) — shape (3,) 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

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, 3 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 0 the same month, 0 later) and 0 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
generate_psi 1 other paper
1 dated earlier
1910.04365 (2019-10)
Asking Easy Questions: A User-Friendly Approach to Active Reward Learning
information_objective_psi 1 other paper
1 dated earlier
1910.04365 (2019-10)
Asking Easy Questions: A User-Friendly Approach to Active Reward Learning
volume_objective_psi 1 other paper
1 dated earlier
1910.04365 (2019-10)
Asking Easy Questions: A User-Friendly Approach to Active Reward Learning

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 2026-11-21.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

As humans interact with autonomous agents to perform increasingly complicated, potentially risky tasks, it is important to be able to efficiently evaluate an agent's performance and correctness. In this paper we formalize and theoretically analyze the problem of efficient value alignment verification: how to efficiently test whether the behavior of another agent is aligned with a human's values. The goal is to construct a kind of "driver's test" that a human can give to any agent which will verify value alignment via a minimal number of queries. We study alignment verification problems with both idealized humans that have an explicit reward function as well as problems where they have implicit values. We analyze verification of exact value alignment for rational agents and propose and analyze heuristic and approximate value alignment verification tests in a wide range of gridworlds and a continuous autonomous driving domain. Finally, we prove that there exist sufficient conditions such that we can verify exact and approximate alignment across an infinite set of test environments via a constantquery-complexity alignment test.

For agents

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

get_harvested_code_for_paper("2012.01557")
get_code_for_paper("2012.01557")
have("2012.01557")

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

curl https://syntology.ai/api/ran/2012.01557.json

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

[![Syntology run record](https://syntology.ai/api/ran/2012.01557.svg)](https://syntology.ai/paper/2012.01557)

Connect an agent — have() is free.