SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2406.03997 · 2024

HackAtari: Atari Learning Environments for Robust and Continual Reinforcement Learning

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 2026-11-22; 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/2406.03997.json · badge: /api/ran/2406.03997.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
k4ntz/HackAtari canonical (classifier label)
classifier basis: census
3 ran · 0 no recorded run
FunctionStatusWhere it lives
modif_funcs Ran
ran; our contract draft was wrong, not the code
k4ntz/HackAtari/hackatari/games/alien.py
code served (permissive licence) · get_code("58dbe0d761ff8506")
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 random

class MockEnv:
    """Mock game environment for testing GameModifications."""
    
    def __init__(self):
        # Initialize RAM with 128 bytes (typical for Atari)
        self.ram = bytearray(128)
        # Set some initial values to make the environment realistic
        self.ram[0] = 1
        self.ram[11] = 50
        self.ram[64] = 3
        self.ram[65] = 0
        self.ram[90] = 0
        self.ram[116] = 100
    
    def get_ram(self):
        """Return a copy of the RAM."""
        return bytearray(self.ram)
    
    def set_ram(self, index, value):
        """Set a RAM value at the given index."""
        self.ram[index] = value & 0xFF  # Ensure value fits in a byte


# Create mock environment
env = MockEnv()

# Create list of active modifications
active_modifs = [
    'last_egg',
    'unlimited_fuel',
    'unlimited_lives',
    'no_enemies',
    'one_alien'
]

# Define ARGS dictionary
ARGS = {
    'env': env,
    'active_modifs': active_modifs
}
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 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
modif_funcs Ran
ran; our contract draft was wrong, not the code
k4ntz/HackAtari/hackatari/games/amidar.py
code served (permissive licence) · get_code("0fc91e791cb7dbea")
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 numpy as np

class MockEnvironment:
    """Mock game environment with RAM storage."""
    
    def __init__(self):
        # Initialize with 256 bytes of RAM (typical for retro games)
        self.ram = np.zeros(256, dtype=np.uint8)
        # Set some initial values in the enemy/player regions (addresses 73-79)
        self.ram[73:80] = np.array([32, 0, 32, 0, 32, 0, 32], dtype=np.uint8)
        # Set lives register
        self.ram[86] = 5
    
    def get_ram(self):
        """Return current RAM state."""
        return self.ram.copy()
    
    def set_ram(self, address, value):
        """Set RAM at given address."""
        self.ram[address] = value & 0xFF  # Ensure 8-bit value

# Create mock environment
mock_env = MockEnvironment()

# Define active modifications list
active_modifs = ['pig_enemies', 'unlimited_lives']

# Create arguments dictionary
ARGS = {
    'env': mock_env,
    'active_modifs': active_modifs
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, 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
modif_funcs Ran
ran; our contract draft was wrong, not the code
k4ntz/HackAtari/hackatari/games/asterix.py
code served (permissive licence) · get_code("ccaded47c78e3540")
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 numpy as np


class MockEnv:
    """Mock environment for testing GameModifications."""
    
    def __init__(self):
        self.ram = np.zeros(256, dtype=np.uint8)
    
    def set_ram(self, address, value):
        """Set a value at a specific RAM address."""
        if address < len(self.ram):
            self.ram[address] = value


# Create a mock environment
env = MockEnv()

# Define active modifications - test with some step modifications
active_modifs = ['obelix', 'set_consumable_1', 'unlimited_lives', 'odd_lines_free']

ARGS = {
    'env': env,
    'active_modifs': active_modifs
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, 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

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 1 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 1 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

Artificial agents' adaptability to novelty and alignment with intended behavior is crucial for their effective deployment. Reinforcement learning (RL) leverages novelty as a means of exploration, yet agents often struggle to handle novel situations, hindering generalization. To address these issues, we propose HackAtari, a framework introducing controlled novelty to the most common RL benchmark, the Atari Learning Environment. HackAtari allows us to create novel game scenarios (including simplification for curriculum learning), to swap the game elements' colors, as well as to introduce different reward signals for the agent. We demonstrate that current agents trained on the original environments include robustness failures, and evaluate HackAtari's efficacy in enhancing RL agents' robustness and aligning behavior through experiments using C51 and PPO. Overall, HackAtari can be used to improve the robustness of current and future RL algorithms, allowing Neuro-Symbolic RL, curriculum RL, causal RL, as well as LLM-driven RL. Our work underscores the significance of developing interpretable in RL agents.

For agents

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

get_harvested_code_for_paper("2406.03997")
get_code_for_paper("2406.03997")
have("2406.03997")

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

curl https://syntology.ai/api/ran/2406.03997.json

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

[![Syntology run record](https://syntology.ai/api/ran/2406.03997.svg)](https://syntology.ai/paper/2406.03997)

Connect an agent — have() is free.