SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2602.15273 · 2026

FrameRef: A Framing Dataset and Simulation Testbed for Modeling Bounded Rational Information Health

Grace Yang, Jiqun Liu, Victor De Lima

arXiv · PDF · Open in the Atlas

Code that ran

8 of 9 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; 2 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 6 failures of our instrument; 1 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/2602.15273.json · badge: /api/ran/2602.15273.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
infosenselab/frameref canonical (classifier label)
classifier basis: census
8 ran · 1 no recorded run
FunctionStatusWhere it lives
build_rewrite_prompt Ran
ran; our contract draft was wrong, not the code
infosenselab/frameref/information_health/dataset/2_restate_claims.py
code served (permissive licence) · get_code("404ceb52369e7a92")
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 str. 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

torch.manual_seed(0)

ARGS = {
    "claim": "The government implemented new climate policies last year.",
    "framing": {
        "description": "Focus on the environmental impact and long-term sustainability goals of the policies.",
        "constraints": [
            "Emphasize scientific consensus",
            "Highlight measurable outcomes",
            "Avoid political attribution"
        ]
    }
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not 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
build_verification_prompt Ran
ran; our contract draft was wrong, not the code
infosenselab/frameref/information_health/dataset/2_restate_claims.py
code served (permissive licence) · get_code("872335c80239106c")
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 str. 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

torch.manual_seed(0)

ARGS = {
    "claim": "Climate change is caused by human activities.",
    "restated_claim": "Human activities, particularly the emission of greenhouse gases, are the primary driver of climate change observed over the past century.",
    "framing_type": "causal"
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not 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
compute_rt Ran
ran, honoured the contract we drafted
infosenselab/frameref/information_health/evaluation/trajectory_model_evaluate.py
code served (permissive licence) · get_code("48a4dad4d33a8c1a")
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 float, 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 torch
import numpy as np

torch.manual_seed(0)

# Build the arguments for compute_rt function
predicted_label = 0
true_label = 0
scores = {
    0: {'word_cond_prob': 0.85},
    1: {'word_cond_prob': 0.42},
    2: {'word_cond_prob': 0.73},
}

ARGS = {
    'predicted_label': predicted_label,
    'true_label': true_label,
    'scores': scores,
}
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 float, 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
decode_first_word Ran
ran; our contract draft was wrong, not the code
infosenselab/frameref/information_health/evaluation/eval_funcs.py
code served (permissive licence) · get_code("8a08cffe4bcd38d6")
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 str. 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 string

class SimpleTokenizer:
    def decode(self, token_ids, skip_special_tokens=False):
        # Simple mock tokenizer that returns different strings based on token id
        decode_map = {
            0: "Hello",
            1: " ",
            2: "world",
            3: ".",
            4: "!",
            5: "test",
        }
        if len(token_ids) == 1:
            tok_id = token_ids[0]
            return decode_map.get(tok_id, f"token_{tok_id}")
        return ""

# Create a simple tokenizer instance
tokenizer = SimpleTokenizer()

# Create token ids that will form a word
# ids = [0, 1] will decode to "Hello" + " " and stop at the space
ids = [0, 1]

ARGS = {
    "ids": ids,
    "tokenizer": tokenizer,
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not 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
find_last_subsequence Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
infosenselab/frameref/information_health/experiments/supervised_finetuning.py
code served (permissive licence) · get_code("2a63432f2dadba71")
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 int, ndim 0, dtype family int, 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 torch

torch.manual_seed(0)

# The function expects:
# - seq: a torch tensor (or similar object with .tolist() method and slicing)
# - subseq: a list or sequence that can be compared with .tolist() output

# Create a small tensor sequence
seq = torch.tensor([1, 2, 3, 4, 5, 2, 3, 4, 6])

# Create a subsequence to search for (as a list for comparison)
subseq = [2, 3, 4]

ARGS = {
    "seq": seq,
    "subseq": subseq,
}
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 int, ndim 0, dtype family int, 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_base_claim_id Ran
ran; our contract draft was wrong, not the code
infosenselab/frameref/information_health/evaluation/trajectory_model_evaluate.py
code served (permissive licence) · get_code("2a756afbefae7fd3")
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 str. 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

torch.manual_seed(0)

ARGS = {
    "claim_id": "base_claim:subclaim:extra"
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not 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
get_claim_text Ran
ran; our contract draft was wrong, not the code
infosenselab/frameref/information_health/evaluation/eval_funcs.py
code served (permissive licence) · get_code("8232385e2fd7eeb6")
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 str. 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

# Define ALLOWED_FRAMINGS - a set of valid framing types
ALLOWED_FRAMINGS = {'factual', 'opinion', 'mixed'}

# Define a valid row dict with all required fields
row = {
    'claim_id': 'claim_001',
    'framing_type': 'factual',
    'restated_claim': 'This is a restated claim text.'
}

# Construct the ARGS dict
ARGS = {
    'row': row,
    'ALLOWED_FRAMINGS': ALLOWED_FRAMINGS
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not 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
policy_loss Ran
ran, honoured the contract we drafted
infosenselab/frameref/information_health/experiments/supervised_finetuning.py
code served (permissive licence) · get_code("0911766220907f3c")
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 Tensor, 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 torch
import torch.nn as nn

torch.manual_seed(0)

# token_weights: used in element-wise multiplication with label_log_probs
# Should be a tensor of shape (batch_size, seq_len) or similar
token_weights = torch.randn(2, 4)

# label_log_probs: used in element-wise multiplication with token_weights
# Should be log probabilities, same shape as token_weights
label_log_probs = torch.log(torch.softmax(torch.randn(2, 4), dim=-1))

ARGS = {
    "token_weights": token_weights,
    "label_log_probs": label_log_probs,
}
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 Tensor, 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
process_batch Not yet run
unverified (harvested, no recorded run)
infosenselab/frameref/information_health/evaluation/eval_funcs.py
code served (permissive licence) · get_code("f46651ffd0ab067b")
How it ran

no invocation recorded.

Also present in other papers' code

Of 9 samples lifted from this paper's linked repositories, 9 carry a function identity and 0 do not. Of the 9 distinct functions with an identity, 2 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 2 later) and 7 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
find_last_subsequence 1 other paper
1 later
2604.10968 (2026-04)
YIELD: A Large-Scale Dataset and Evaluation Framework for Information Elicitation Agents
policy_loss 1 other paper
1 later
2604.10968 (2026-04)
YIELD: A Large-Scale Dataset and Evaluation Framework for Information Elicitation 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

Information ecosystems increasingly shape how people internalize exposure to adverse digital experiences, raising concerns about the long-term consequences for information health. In modern search and recommendation systems, ranking and personalization policies play a central role in shaping such exposure and its longterm effects on users. To study these effects in a controlled setting, we present FrameRef, a large-scale dataset of 1,073,740 systematically reframed claims across five framing dimensions: authoritative, consensus, emotional, prestige, and sensationalist, and propose a simulation-based framework for modeling sequential information exposure and reinforcement dynamics characteristic of ranking and recommendation systems. Within this framework, we construct framing-sensitive agent personas by fine-tuning language models with framing-conditioned loss attenuation, inducing targeted biases while preserving overall task competence. Using Monte Carlo trajectory sampling, we show that small, systematic shifts in acceptance and confidence can compound over time, producing substantial divergence in cumulative information health trajectories. Human evaluation further confirms that FrameRef's generated framings measurably affect human judgment. Together, our dataset and framework provide a foundation for systematic information health research through simulation, complementing and informing responsible human-centered research. We release FrameRef, code, documentation, human evaluation data, and persona adapter models at https://github.com/infosenselab/frameref.

For agents

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

get_harvested_code_for_paper("2602.15273")
get_code_for_paper("2602.15273")
have("2602.15273")

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

curl https://syntology.ai/api/ran/2602.15273.json

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

[![Syntology run record](https://syntology.ai/api/ran/2602.15273.svg)](https://syntology.ai/paper/2602.15273)

Connect an agent — have() is free.