SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2510.14967 · ICLR · 2025

Information Gain-based Policy Optimization: A Simple and Effective Approach for Multi-Turn Search Agents

Wei Yao, Guoqing Wang, Sunhao Dai, Zeyu Gan, Guangze Ye, Yong Deng, Xiaofeng Wu, Zhenzhe Ying

arXiv · PDF · Open in the Atlas

Code that ran

8 of 17 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, 2 violated it and 0 have no contract verdict; 6 failures of our instrument; 9 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/2510.14967.json · badge: /api/ran/2510.14967.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
GuoqingWang1/IGPO canonical (classifier label)
classifier basis: census
8 ran · 9 no recorded run
FunctionStatusWhere it lives
check_tags_balance Ran
ran, violated the contract we drafted
GuoqingWang1/IGPO/evaluate/cacluate_metrics.py
code served (permissive licence) · get_code("54d0cd89e3fa4317")
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.

# Setup code for check_tags_balance function

ARGS = {
    "solution_str": "<tool_call>function()</tool_call><think>reasoning</think><answer>result</answer>"
}
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
get_tool_names Ran
ran; our contract draft was wrong, not the code
GuoqingWang1/IGPO/tools_server/tools.py
code served (permissive licence) · get_code("3ef330f50eba38f6")
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 Dict, Any

torch.manual_seed(0)

# The function get_tool_names takes a config parameter (Dict[str, Any])
# and returns a list of tool names. Looking at the function body,
# it doesn't actually use the config parameter - it just returns a hardcoded list.
# So we can pass None or an empty dict or any dict.

ARGS = {
    'config': None
}
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
get_tools Ran
ran; our contract draft was wrong, not the code
GuoqingWang1/IGPO/tools_server/tools.py
code served (permissive licence) · get_code("c5677288a38d7a82")
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 dict. 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
from typing import Dict, Any

# The function get_tools takes one optional parameter: config
# Looking at the function body, it doesn't actually use the config parameter
# It just returns a hardcoded dictionary with the WEB_SEARCH_TOOL

# config is optional (defaults to None), so we can pass None or a dict
# Let's pass None since the function doesn't use it

ARGS = {
    'config': None
}
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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict 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
is_ipv4 Ran
ran, violated the contract we drafted
GuoqingWang1/IGPO/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_text Ran
ran; our contract draft was wrong, not the code
GuoqingWang1/IGPO/evaluate/cacluate_metrics.py
code served (permissive licence) · get_code("567311eaeb3eeea6")
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 re
import string

# Construct valid arguments for preprocess_text function
text_input = "Hello, World! This is a TEST. Multiple   spaces here!!!"

ARGS = {
    "text": text_input
}
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
string_to_uuid Ran
ran; our contract draft was wrong, not the code
GuoqingWang1/IGPO/tools_server/util.py
code served (permissive licence) · get_code("8f70e95490c5a263")
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 uuid

def string_to_uuid(input_string: str) -> str:
    """Convert string to deterministic UUID."""
    return str(uuid.uuid5(uuid.NAMESPACE_DNS, str(input_string)))

ARGS = {
    "input_string": "example_string_for_uuid"
}
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
union_tensor_dict Ran
ran; our contract draft was wrong, not the code
GuoqingWang1/IGPO/verl/protocol.py
code served (permissive licence) · get_code("21331a58f93375e4")
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 TensorDict. 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
from tensordict import TensorDict

torch.manual_seed(0)

# Create two TensorDicts with identical batch size
batch_size = torch.Size([2])

# First tensor dict with some keys
tensor_dict1 = TensorDict(
    {
        "shared_key": torch.randn(2, 4),
        "key1": torch.randn(2, 3),
    },
    batch_size=batch_size,
)

# Second tensor dict with overlapping and new keys
tensor_dict2 = TensorDict(
    {
        "shared_key": tensor_dict1["shared_key"].clone(),  # Same tensor values
        "key2": torch.randn(2, 5),
    },
    batch_size=batch_size,
)

ARGS = {
    "tensor_dict1": tensor_dict1,
    "tensor_dict2": tensor_dict2,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
tensordict, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type TensorDict — a TensorDict 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
unpad_dataproto Ran
ran; our contract draft was wrong, not the code
GuoqingWang1/IGPO/verl/protocol.py
code served (permissive licence) · get_code("25f0ea3f460f6ce1")
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 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 class for DataProto that supports slicing."""
    def __init__(self, data):
        self.data = data
    
    def __getitem__(self, key):
        """Support slicing to return a new DataProto with sliced data."""
        sliced_data = self.data[key]
        return DataProto(sliced_data)
    
    def __repr__(self):
        return f"DataProto({self.data})"

# Create a small DataProto instance with some data
data_tensor = torch.randn(2, 4)
data_proto = DataProto(data_tensor)

ARGS = {
    "data": data_proto,
    "pad_size": 0
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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
call_gpt_4o_mini Not yet run
unverified (harvested, no recorded run)
GuoqingWang1/IGPO/evaluate/cacluate_metrics.py
code served (permissive licence) · get_code("e5ead6aad013dbb3")
How it ran

no invocation recorded.

get_generation_config Not yet run
unverified (harvested, no recorded run)
GuoqingWang1/IGPO/verl/utils/model.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from shanjf666/CoCoV, is pointer only (licence: NONE) · get_code("68aa34a57b353cae")
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 shanjf666/CoCoV, 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. No invocation recorded.

get_huggingface_actor_config Not yet run
unverified (harvested, no recorded run)
GuoqingWang1/IGPO/verl/utils/model.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from shanjf666/CoCoV, is pointer only (licence: NONE) · get_code("9411d1fb7e4f68e5")
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 shanjf666/CoCoV, 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. No invocation recorded.

get_weight_loader Not yet run
unverified (harvested, no recorded run)
GuoqingWang1/IGPO/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)
GuoqingWang1/IGPO/verl/models/weight_loader_registry.py
code served (permissive licence) · get_code("1008c22d29f85015")
How it ran

no invocation recorded.

initialize_system_prompt Not yet run
unverified (harvested, no recorded run)
GuoqingWang1/IGPO/tools_server/initialize_prompts.py
code served (permissive licence) · get_code("3342a992efb2b01e")
How it ran

no invocation recorded.

is_ipv6 Not yet run
unverified (harvested, no recorded run)
GuoqingWang1/IGPO/verl/utils/net_utils.py
code served (permissive licence) · get_code("705afe175ad8e404")
How it ran

no invocation recorded.

populate_template Not yet run
unverified (harvested, no recorded run)
GuoqingWang1/IGPO/tools_server/initialize_prompts.py
code served (permissive licence) · get_code("95d41d682e3b2f44")
How it ran

no invocation recorded.

squeeze Not yet run
unverified (harvested, no recorded run)
GuoqingWang1/IGPO/verl/utils/model.py
code served (permissive licence) · get_code("3b15e2ac7497c441")
How it ran

no invocation recorded.

Also present in other papers' code

Of 17 samples lifted from this paper's linked repositories, 17 carry a function identity and 0 do not. Of the 17 distinct functions with an identity, 9 are also present in other papers whose code we hold (9 with at least one paper dated earlier, 8 the same month, 8 later) and 8 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
unpad_dataproto 89 other papers
27 dated earlier, 1 the same month, 57 later, 1 same year, order not determinable, 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 86 more not named here
squeeze 82 other papers
26 dated earlier, 1 the same month, 51 later, 1 same year, order not determinable, 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
union_tensor_dict 64 other papers
30 dated earlier, 1 the same month, 29 later, 1 same year, order not determinable, 3 undated · 4 only as an assembly brick
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 61 more not named here
get_generation_config 63 other papers
15 dated earlier, 1 the same month, 45 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
6 dated earlier, 1 the same month, 49 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
25 dated earlier, 1 the same month, 24 later, 1 same year, order not determinable, 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
is_ipv4 41 other papers
3 dated earlier, 1 the same month, 35 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
3 dated earlier, 1 the same month, 35 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
get_weight_saver 2 other papers
2 dated earlier
2505.14362 (2025-05)
DeepEyes: Incentivizing "Thinking with Images" via Reinforcement Learning
2506.08672 (2025-06)
RuleReasoner: Reinforced Rule-based Reasoning via Domain-aware Dynamic Sampling

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

Large language model (LLM)-based agents are increasingly trained with reinforcement learning (RL) to enhance their ability to interact with external environments through tool use, particularly in search-based settings that require multi-turn reasoning and knowledge acquisition. However, existing approaches typically rely on outcome-based rewards that are only provided exclusively upon generating the final answer. This reward sparsity becomes particularly problematic in multi-turn settings, where long trajectories exacerbate three critical issues: (i) advantage collapse, where all rollouts receive identical rewards and provide no useful learning signals; (ii) lack of fine-grained credit assignment, where the correctness of intermediate turns is obscured, especially in long-horizon tasks; and (iii) poor sample efficiency, where each rollout yields only a single outcome signal, leading to low data utilization. In this paper, we propose Information Gain-based Policy Optimization (IGPO), a simple yet effective RL framework that provides dense and intrinsic supervision for multi-turn agent training. IGPO models each interaction turn as an incremental process of acquiring information about the ground truth, and defines turn-level rewards as the marginal increase in the policy's probability of producing the correct answer. Unlike prior process-level reward approaches that depend on external reward models or costly Monte Carlo estimation, IGPO derives intrinsic rewards directly from the model's own belief updates. These intrinsic turn-level rewards are combined with outcome-level supervision to form dense reward signals. Extensive experiments on both in-domain and out-of-domain benchmarks demonstrate that IGPO consistently outperforms strong baselines in multiturn scenarios, achieving higher accuracy and improved data efficiency. Our code is available at https://github.com/GuoqingWang1/IGPO.

For agents

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

get_harvested_code_for_paper("2510.14967")
get_code_for_paper("2510.14967")
have("2510.14967")

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

curl https://syntology.ai/api/ran/2510.14967.json

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

[![Syntology run record](https://syntology.ai/api/ran/2510.14967.svg)](https://syntology.ai/paper/2510.14967)

Connect an agent — have() is free.