SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2601.04786 · 2026

AgentOCR: Reimagining Agent History via Optical Self-Compression

Feng Chen, Ming Yan, Bo An, Xin Cheng, Haiyang Xu, Lang Feng, Zhenglin Wan, Fuchao Yang

arXiv · PDF · Open in the Atlas

Code that ran

14 of 20 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; 1 honoured the contract we drafted, 2 violated it and 0 have no contract verdict; 11 failures of our instrument; 6 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/2601.04786.json · badge: /api/ran/2601.04786.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
langfengQ/AgentOCR canonical (classifier label)
classifier basis: census
14 ran · 6 no recorded run
FunctionStatusWhere it lives
apply_compact_mode Ran
ran; our contract draft was wrong, not the code
langfengQ/AgentOCR/agentocr/utils.py
code served (permissive licence) · get_code("527b22fba182380a")
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

COMPACT_NEWLINE_SYMBOL = '⏎'

def apply_compact_mode(text: str, symbol: str=COMPACT_NEWLINE_SYMBOL) -> str:
    """
    Convert text to compact mode by replacing newlines with a special symbol.
    
    Args:
        text: Original text with newlines
        symbol: Symbol to replace newlines with (default: ⏎)
    
    Returns:
        Text with newlines replaced by the symbol, treated as single paragraph
    """
    if not text:
        return ''
    compact_text = text.replace('\n', f' {symbol} ')
    compact_text = ' '.join(compact_text.split())
    return compact_text


ARGS = {
    'text': 'Hello\nWorld\nThis\nis\nmultiline\ntext',
    'symbol': '⏎'
}
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
are_similar Ran
ran, violated the contract we drafted
langfengQ/AgentOCR/gigpo/core_gigpo.py
code served (permissive licence) · get_code("0588eb70d1fa7310")
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 bool, 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.

from difflib import SequenceMatcher

ARGS = {
    "a": "hello world",
    "b": "hello world",
    "threshold": 0.95
}
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 bool, 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
compute_ce_dpo_loss_rm Ran
ran; our contract draft was wrong, not the code
langfengQ/AgentOCR/recipe/prime/prime_core_algos.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from LeapLabTHU/OPDVR, is pointer only (licence: NONE) · get_code("f220595adfc5e83c")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from LeapLabTHU/OPDVR, and GitHub detected no licence in that repository 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.

compute_detach_dpo_loss_rm Ran
ran; our contract draft was wrong, not the code
langfengQ/AgentOCR/recipe/prime/prime_core_algos.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from LeapLabTHU/OPDVR, is pointer only (licence: NONE) · get_code("4d91e4383716e42c")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from LeapLabTHU/OPDVR, and GitHub detected no licence in that repository 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.

compute_online_dpo_loss Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
langfengQ/AgentOCR/recipe/spin/core_algos.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from LeapLabTHU/OPDVR, is pointer only (licence: NONE) · get_code("f19825983a523003")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from LeapLabTHU/OPDVR, and GitHub detected no licence in that repository 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.

compute_onlinedpo_pref Ran
ran; our contract draft was wrong, not the code
langfengQ/AgentOCR/recipe/spin/core_algos.py
code served (permissive licence) · get_code("7a988deb09f43a64")
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 1, dtype family float, shape (4,) 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)

# Create valid arguments for compute_onlinedpo_pref
# The function expects:
# - token_level_rewards: shape [batch_size * 2, seq_len]
# - response_mask: shape [batch_size * 2, seq_len]

batch_size = 2  # number of pairs
seq_len = 4

# Create token_level_rewards with shape [4, 4] (batch_size * 2 = 4, seq_len = 4)
token_level_rewards = torch.randn(batch_size * 2, seq_len)

# Create response_mask with same shape [4, 4]
# Binary mask where 1 indicates valid tokens, 0 indicates padding
response_mask = torch.ones(batch_size * 2, seq_len, dtype=torch.float32)
# Make some positions masked out for variety
response_mask[0, 3] = 0
response_mask[2, 2:] = 0

ARGS = {
    "token_level_rewards": token_level_rewards,
    "response_mask": response_mask,
}
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 1, dtype family float, shape (4,) — shape (4,) 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
extract_dataproto_via_active_mask Ran
ran; our contract draft was wrong, not the code
langfengQ/AgentOCR/verl/protocol.py
code served (permissive licence) · get_code("90d951e48e961af1")
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 DataProto. 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)

class DataProto:
    """Stand-in DataProto class that supports select_idxs and len operations."""
    
    def __init__(self, data, size=5):
        self.data = data
        self.size = size
    
    def __len__(self):
        return self.size
    
    def select_idxs(self, idxs):
        """Select subset of data by indices."""
        new_size = len(idxs)
        return DataProto(self.data[idxs] if isinstance(self.data, np.ndarray) else self.data, size=new_size)


batch_input = DataProto(
    data=np.random.randn(5, 4),
    size=5
)

active_mask_list = [True, False, True, True, False]

ARGS = {
    "batch_input": batch_input,
    "active_mask": active_mask_list,
}
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 DataProto — a DataProto output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DataProto 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_compact_symbol_positions Ran
ran, honoured the contract we drafted
langfengQ/AgentOCR/agentocr/utils.py
code served (permissive licence) · get_code("7e9b7a33e55bea26")
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 list, ndim 1, dtype family int; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim 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)

# Construct valid arguments for get_compact_symbol_positions
# The function takes:
# - text: str (required)
# - symbol: str (optional, defaults to COMPACT_NEWLINE_SYMBOL)

# Create test text with multiple instances of the compact newline symbol
COMPACT_NEWLINE_SYMBOL = '⏎'

text_with_symbols = f"Hello{COMPACT_NEWLINE_SYMBOL}World{COMPACT_NEWLINE_SYMBOL}Test{COMPACT_NEWLINE_SYMBOL}String"
custom_symbol = "⏎"

ARGS = {
    "text": text_with_symbols,
    "symbol": custom_symbol
}
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 list, ndim 1, dtype family int — a list output has no array shape of its own; none is stated
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_kl_controller Ran
ran; our contract draft was wrong, not the code
langfengQ/AgentOCR/recipe/spin/core_algos.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from LeapLabTHU/OPDVR, is pointer only (licence: NONE) · get_code("757d109d90121142")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from LeapLabTHU/OPDVR, and GitHub detected no licence in that repository 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.

is_ipv4 Ran
ran, violated the contract we drafted
langfengQ/AgentOCR/verl/utils/net_utils.py
code served (permissive licence) · get_code("735e93af30b15fa9")
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 bool, 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 ipaddress

ARGS = {
    "ip_str": "192.168.1.1"
}
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 bool, 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
preprocess_trajectory_contexts Ran
ran; our contract draft was wrong, not the code
langfengQ/AgentOCR/agentocr/utils.py
code served (permissive licence) · get_code("e43e286f03310f0c")
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 list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim 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
from typing import List

# Set seed for reproducibility
torch.manual_seed(0)

# Construct the trajectory_contexts parameter
# The function expects a list of strings where escaped quotes need to be replaced
trajectory_contexts = [
    'context with \\"escaped\\" quotes',
    'another context with \\"quote\\"',
    'simple context without escapes',
    'mixed \\"escaped\\" and normal quotes'
]

ARGS = {
    'trajectory_contexts': trajectory_contexts
}
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 list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
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
reward_func Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
langfengQ/AgentOCR/recipe/r1/reward_score.py
code served (permissive licence) · get_code("c30b2a96103abd2c")
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 MagicMock. 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)

# Mock the recipe.r1.tasks modules to avoid import errors
import sys
from unittest.mock import MagicMock

mock_math = MagicMock()
mock_math.compute_score = lambda sol, gt: 0.85

mock_gpqa = MagicMock()
mock_gpqa.compute_score = lambda sol, gt: 0.92

mock_livecodebench = MagicMock()
mock_livecodebench.compute_score = lambda sol, gt: 0.78

sys.modules['recipe'] = MagicMock()
sys.modules['recipe.r1'] = MagicMock()
sys.modules['recipe.r1.tasks'] = MagicMock()
sys.modules['recipe.r1.tasks.math'] = mock_math
sys.modules['recipe.r1.tasks.gpqa'] = mock_gpqa
sys.modules['recipe.r1.tasks.livecodebench'] = mock_livecodebench

# Define arguments for reward_func
ARGS = {
    'data_source': 'Maxwell-Jia/AIME_2024',
    'solution_str': 'The answer is 42.',
    'ground_truth': '42',
    'extra_info': None
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, recipe, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type MagicMock — a MagicMock output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a MagicMock output the recorded ndim 1 and dtype family None 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
split_into_segments Ran
ran; our contract draft was wrong, not the code
langfengQ/AgentOCR/agentocr/ocrtool.py
code served (permissive licence) · get_code("6c8217f43840eab6")
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 list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim 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

# The function takes a single parameter: history (a string)
# It splits the history by newlines and strips whitespace from each line
# Returns a list of non-empty line segments

ARGS = {
    "history": "Line 1\nLine 2\nLine 3\n\nLine 4"
}
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 list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
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
to_hashable Ran
ran; our contract draft was wrong, not the code
langfengQ/AgentOCR/gigpo/core_gigpo.py
code served (permissive licence) · get_code("862de021e4c430ca")
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 numpy as np

ARGS = {
    "x": [
        42,
        3.14,
        "hello",
        True,
        np.int64(10),
        np.float32(2.5),
        np.array([1, 2, 3, 4]),
        [1, 2, "test"],
        (3.14, "tuple", True),
        {"key1": 5, "key2": "value", "nested": [1, 2, 3]},
        np.array([[1, 2], [3, 4]]),
        {"a": np.int32(1), "b": np.array([5, 6, 7])},
        [np.float64(1.5), {"inner": "dict"}, (1, 2, 3)],
    ]
}
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 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
get_generation_config Not yet run
unverified (harvested, no recorded run)
langfengQ/AgentOCR/verl/utils/model.py
code served (permissive licence) · get_code("5c059b45527bdff7")
How it ran

no invocation recorded.

get_huggingface_actor_config Not yet run
unverified (harvested, no recorded run)
langfengQ/AgentOCR/verl/utils/model.py
code served (permissive licence) · get_code("0a5b4c20bc4ff94b")
How it ran

no invocation recorded.

get_weight_loader Not yet run
unverified (harvested, no recorded run)
langfengQ/AgentOCR/verl/models/weight_loader_registry.py
code served (permissive licence) · get_code("acf240079d434746")
How it ran

no invocation recorded.

get_weight_saver Not yet run
unverified (harvested, no recorded run)
langfengQ/AgentOCR/verl/models/weight_loader_registry.py
code served (permissive licence) · get_code("55f217ce71432b58")
How it ran

no invocation recorded.

is_ipv6 Not yet run
unverified (harvested, no recorded run)
langfengQ/AgentOCR/verl/utils/net_utils.py
code served (permissive licence) · get_code("705afe175ad8e404")
How it ran

no invocation recorded.

squeeze Not yet run
unverified (harvested, no recorded run)
langfengQ/AgentOCR/verl/utils/model.py
code served (permissive licence) · get_code("3b15e2ac7497c441")
How it ran

no invocation recorded.

Also present in other papers' code

Of 20 samples lifted from this paper's linked repositories, 20 carry a function identity and 0 do not. Of the 20 distinct functions with an identity, 15 are also present in other papers whose code we hold (15 with at least one paper dated earlier, 14 the same month, 15 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
squeeze 82 other papers
32 dated earlier, 3 the same month, 44 later, 3 undated
2409.19256 (2024-09)
HybridFlow: A Flexible and Efficient RLHF Framework
2410.21236 (2024-10)
Flaming-hot Initiation with Regular Execution Sampling for Large Language Models
2025.findings-emnlp.1015 (2025 (year only))
MT-R1-Zero: Advancing LLM-based Machine Translation via R1-Zero-like Reinforcement Learning
and 79 more not named here
get_generation_config 63 other papers
19 dated earlier, 2 the same month, 40 later, 2 undated
2409.19256 (2024-09)
HybridFlow: A Flexible and Efficient RLHF Framework
2410.21236 (2024-10)
Flaming-hot Initiation with Regular Execution Sampling for Large Language Models
2502.19613 (2025-02)
Self-rewarding correction for mathematical reasoning
and 60 more not named here
get_weight_loader 58 other papers
10 dated earlier, 4 the same month, 42 later, 2 undated
2409.19256 (2024-09)
HybridFlow: A Flexible and Efficient RLHF Framework
2505.14362 (2025-05)
DeepEyes: Incentivizing "Thinking with Images" via Reinforcement Learning
2505.20046 (2025-05)
REARANK: Reasoning Re-ranking Agent via Reinforcement Learning
and 55 more not named here
get_huggingface_actor_config 53 other papers
31 dated earlier, 2 the same month, 18 later, 2 undated
2409.19256 (2024-09)
HybridFlow: A Flexible and Efficient RLHF Framework
2410.21236 (2024-10)
Flaming-hot Initiation with Regular Execution Sampling for Large Language Models
2025.findings-emnlp.1015 (2025 (year only))
MT-R1-Zero: Advancing LLM-based Machine Translation via R1-Zero-like Reinforcement Learning
and 50 more not named here
compute_ce_dpo_loss_rm 46 other papers
11 dated earlier, 4 the same month, 28 later, 3 undated
2409.19256 (2024-09)
HybridFlow: A Flexible and Efficient RLHF Framework
2410.02992 (2024-10)
Learning to Better Search with Language Models via Guided Reinforced Self-Training
2504.16084 (2025-04)
TTRL: Test-Time Reinforcement Learning
and 43 more not named here
compute_detach_dpo_loss_rm 44 other papers
9 dated earlier, 4 the same month, 28 later, 3 undated
2409.19256 (2024-09)
HybridFlow: A Flexible and Efficient RLHF Framework
2410.02992 (2024-10)
Learning to Better Search with Language Models via Guided Reinforced Self-Training
2504.16084 (2025-04)
TTRL: Test-Time Reinforcement Learning
and 41 more not named here
is_ipv4 41 other papers
7 dated earlier, 2 the same month, 30 later, 2 undated
2409.19256 (2024-09)
HybridFlow: A Flexible and Efficient RLHF Framework
2506.08672 (2025-06)
RuleReasoner: Reinforced Rule-based Reasoning via Domain-aware Dynamic Sampling
2506.12935 (2025-06)
SoundMind: RL-Incentivized Logic Reasoning for Audio-Language Models
and 38 more not named here
is_ipv6 41 other papers
7 dated earlier, 2 the same month, 30 later, 2 undated
2409.19256 (2024-09)
HybridFlow: A Flexible and Efficient RLHF Framework
2506.08672 (2025-06)
RuleReasoner: Reinforced Rule-based Reasoning via Domain-aware Dynamic Sampling
2506.12935 (2025-06)
SoundMind: RL-Incentivized Logic Reasoning for Audio-Language Models
and 38 more not named here
compute_online_dpo_loss 38 other papers
7 dated earlier, 3 the same month, 25 later, 3 undated
2409.19256 (2024-09)
HybridFlow: A Flexible and Efficient RLHF Framework
2410.02992 (2024-10)
Learning to Better Search with Language Models via Guided Reinforced Self-Training
2504.16084 (2025-04)
TTRL: Test-Time Reinforcement Learning
and 35 more not named here
get_kl_controller 38 other papers
7 dated earlier, 3 the same month, 25 later, 3 undated
2409.19256 (2024-09)
HybridFlow: A Flexible and Efficient RLHF Framework
2410.02992 (2024-10)
Learning to Better Search with Language Models via Guided Reinforced Self-Training
2504.16084 (2025-04)
TTRL: Test-Time Reinforcement Learning
and 35 more not named here
reward_func 28 other papers
10 dated earlier, 1 the same month, 15 later, 2 undated
2409.19256 (2024-09)
HybridFlow: A Flexible and Efficient RLHF Framework
2504.16084 (2025-04)
TTRL: Test-Time Reinforcement Learning
2505.14362 (2025-05)
DeepEyes: Incentivizing "Thinking with Images" via Reinforcement Learning
and 25 more not named here
get_weight_saver 10 other papers
2 dated earlier, 1 the same month, 6 later, 1 undated
2512.14157 (2025-12)
Ophiuchus: Incentivizing Tool-augmented "Think with Images" for Joint Medical Segmentation, Understanding and Reasoning
2512.16848 (2025-12)
Meta-RL Induces Exploration in Language Agents
2601.04767 (2026-01)
AT 2 PO: Agentic Turn-based Policy Optimization via Tree Search
and 7 more not named here
to_hashable 10 other papers
3 dated earlier, 1 the same month, 6 later · 4 only as an assembly brick
2505.10978 (2025-05)
Group-in-Group Policy Optimization for LLM Agent Training
2509.13347 (2025-09)
OpenHA: A Series of Open-Source Hierarchical Agentic Models in Minecraft
2512.14157 (2025-12)
Ophiuchus: Incentivizing Tool-augmented "Think with Images" for Joint Medical Segmentation, Understanding and Reasoning
and 7 more not named here
compute_onlinedpo_pref 7 other papers
3 dated earlier, 1 the same month, 3 later
2506.12935 (2025-06)
SoundMind: RL-Incentivized Logic Reasoning for Audio-Language Models
2510.04678 (2025-10)
Multi-Agent Tool-Integrated Policy Optimization
2512.14157 (2025-12)
Ophiuchus: Incentivizing Tool-augmented "Think with Images" for Joint Medical Segmentation, Understanding and Reasoning
and 4 more not named here
are_similar 6 other papers
2 dated earlier, 4 later · 2 only as an assembly brick
2505.10978 (2025-05)
Group-in-Group Policy Optimization for LLM Agent Training
2512.14157 (2025-12)
Ophiuchus: Incentivizing Tool-augmented "Think with Images" for Joint Medical Segmentation, Understanding and Reasoning
2604.02268 (2026-04)
SKILL0: In-Context Agentic Reinforcement Learning for Skill Internalization
and 3 more not named here

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

Recent advances in large language models (LLMs) enable agentic systems trained with reinforcement learning (RL) over multi-turn interaction, but practical deployment is bottlenecked by rapidly growing textual histories that inflate token and memory costs. We introduce Agen-tOCR, a framework that exploits visual tokens' superior information density by representing the accumulated observation-action history as a compact rendered image. To make multi-turn rollouts scalable, AgentOCR proposes segment optical caching. By decomposing history into hashable segments and maintaining a visual cache, this mechanism eliminates redundant rerendering. Beyond fixed rendering, AgentOCR introduces agentic self-compression, where the agent actively emits a compression rate and is trained with compression-aware reward to adaptively balance task success and token efficiency. We conduct extensive experiments on challenging agentic benchmarks, ALFWorld and search-based QA. Remarkably, AgentOCR preserves over 95% of text-based agent performance while substantially reducing token consumption (>50%), yielding consistent token and memory efficiency. Further analysis validates a 20× rendering speedup from optical caching and effective self-compression balancing. Our code is available at https: //github.com/langfengQ/AgentOCR.

For agents

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

get_harvested_code_for_paper("2601.04786")
get_code_for_paper("2601.04786")
have("2601.04786")

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

curl https://syntology.ai/api/ran/2601.04786.json

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

[![Syntology run record](https://syntology.ai/api/ran/2601.04786.svg)](https://syntology.ai/paper/2601.04786)

Connect an agent — have() is free.