SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1912.12191 · 2019

Explain Your Move: Understanding Agent Actions Using Specific and Relevant Feature Attribution

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; 3 honoured the contract we drafted, 1 violated it and 0 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/1912.12191.json · badge: /api/ran/1912.12191.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
nikaashpuri/sarfa-saliency canonical (classifier label)
classifier basis: classifier_2025
3 ran · 0 no recorded run
rl-interpretation/understandingRL canonical (classifier label)
classifier basis: classifier_2025
2 ran · 0 no recorded run
copy not recorded community
not in the paper's declared repositories
1 ran · 0 no recorded run
FunctionStatusWhere it lives
computeSaliencyUsingSarfa Ran
ran, honoured the contract we drafted
nikaashpuri/sarfa-saliency/sarfa_saliency.py
pointer only (licence: NONE) · get_code("428b9e2c28356cde")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, nikaashpuri/sarfa-saliency 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.

cross_entropy Ran
ran; our contract draft was wrong, not the code
rl-interpretation/understandingRL/visualize_atari/saliency.py
code served (permissive licence) · get_code("636e79ded1f9a112")
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) and returned float64, 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 numpy as np
from scipy.stats import entropy

# Analyze the function:
# - L_policy: numpy array, indexed with L_idx, sliced with [:L_idx] and [L_idx+1:]
# - l_policy: numpy array, same operations as L_policy
# - L_idx: integer index into the arrays

# Create small numpy arrays with distinct sizes
np.random.seed(0)

# L_policy and l_policy should be 1D arrays with at least 2 elements
# Using size 5 to allow indexing and slicing
L_policy = np.array([0.2, 0.3, 0.25, 0.15, 0.1], dtype=np.float64)
l_policy = np.array([0.25, 0.28, 0.22, 0.18, 0.07], dtype=np.float64)

# L_idx should be a valid index into the arrays (0 to 4)
L_idx = 2

ARGS = {
    "L_policy": L_policy,
    "l_policy": l_policy,
    "L_idx": L_idx,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, scipy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, 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
cross_entropy Ran
ran, honoured the contract we drafted
nikaashpuri/sarfa-saliency/sarfa_saliency.py
pointer only (licence: NONE) · get_code("e3db66ce577ec4b4")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, nikaashpuri/sarfa-saliency 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.

cross_entropy_mcts Ran
ran, honoured the contract we drafted
rl-interpretation/understandingRL/go/go_saliency.py
code served (permissive licence) · get_code("509a45cb8014fde0")
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) and returned float64, 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 numpy as np
import math
from scipy.stats import entropy

# Create sample probability distributions as dictionaries
# dict1 and dict2 map action names to probabilities
dict1 = {
    'action_a': 0.5,
    'action_b': 0.3,
    'action_c': 0.2
}

dict2 = {
    'action_a': 0.4,
    'action_b': 0.35,
    'action_c': 0.25
}

# a_b is an action to exclude from the calculation
a_b = 'action_b'

ARGS = {
    'dict1': dict1,
    'dict2': dict2,
    'a_b': a_b
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, scipy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, 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
get_mask Ran
ran; our contract draft was wrong, not the code
this paper's copy was not recorded; identical code first harvested from greydanus/visualize_atari
pointer only; the stored copy get_code serves, from greydanus/visualize_atari, is pointer only (licence: NONE) · get_code("ef0d62e8fcceba28")
How it ran

The body of this sample is withheld: this paper's copy of the code was not recorded (its link names no repository or source path), so no licence was read for 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.

your_softmax Ran
ran, violated the contract we drafted
nikaashpuri/sarfa-saliency/sarfa_saliency.py
pointer only (licence: NONE) · get_code("aa9aa8dd7b99d120")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, nikaashpuri/sarfa-saliency 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.

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
get_mask 1 other paper
1 dated earlier
1711.00138 (2017-11)
Visualizing and Understanding Atari Agents

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

As deep reinforcement learning (RL) is applied to more tasks, there is a need to visualize and understand the behavior of learned agents. Saliency maps explain agent behavior by highlighting the features of the input state that are most relevant for the agent in taking an action. Existing perturbation-based approaches to compute saliency often highlight regions of the input that are not relevant to the action taken by the agent. Our proposed approach, SARFA (Specific and Relevant Feature Attribution), generates more focused saliency maps by balancing two aspects (specificity and relevance) that capture different desiderata of saliency. The first captures the impact of perturbation on the relative expected reward of the action to be explained. The second downweighs irrelevant features that alter the relative expected rewards of actions other than the action to be explained. We compare SARFA with existing approaches on agents trained to play board games (Chess and Go) and Atari games (Breakout, Pong and Space Invaders). We show through illustrative examples (Chess, Atari, Go), human studies (Chess), and automated evaluation methods (Chess) that SARFA generates saliency maps that are more interpretable for humans than existing approaches. For the code release and demo videos, see https://nikaashpuri.github.io/sarfa-saliency/.

For agents

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

get_harvested_code_for_paper("1912.12191")
get_code_for_paper("1912.12191")
have("1912.12191")

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

curl https://syntology.ai/api/ran/1912.12191.json

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

[![Syntology run record](https://syntology.ai/api/ran/1912.12191.svg)](https://syntology.ai/paper/1912.12191)

Connect an agent — have() is free.