Add four held-out ConsumptionQA examples
Browse files- Model_Card_EN.md +45 -0
- Model_Card_ZH.md +45 -0
- README.md +45 -0
- qa_examples.json +62 -0
Model_Card_EN.md
CHANGED
|
@@ -17,6 +17,51 @@ tags:
|
|
| 17 |
|
| 18 |
**FinL-bilingual** is the bilingual initial-supervision checkpoint, with a multilingual encoder and English/Chinese support. Its MLX runtime is **FinL-bilingual-mlx**. The paper's A100 single-request timing protocol and 22.72× Jev shared-input-token-throughput result belong to this bilingual checkpoint pair. The main English-base six-round checkpoints are [FinL](https://huggingface.co/ya9919/FinL) and [FinL-mlx](https://huggingface.co/ya9919/FinL-mlx).
|
| 19 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 20 |
## Project background
|
| 21 |
|
| 22 |
S²-MAS is a society-scale multi-agent simulation framework for studying how heterogeneous household decisions become population outcomes under changing conditions. It connects a constructed resource of **500 million household personas**, reusable scenarios, trainable decision models and stateful world updates. FinL-bilingual supplies a reusable decision policy: it scores the complete alternatives for a household state and question, allowing the same policy interface to serve different tasks.
|
|
|
|
| 17 |
|
| 18 |
**FinL-bilingual** is the bilingual initial-supervision checkpoint, with a multilingual encoder and English/Chinese support. Its MLX runtime is **FinL-bilingual-mlx**. The paper's A100 single-request timing protocol and 22.72× Jev shared-input-token-throughput result belong to this bilingual checkpoint pair. The main English-base six-round checkpoints are [FinL](https://huggingface.co/ya9919/FinL) and [FinL-mlx](https://huggingface.co/ya9919/FinL-mlx).
|
| 19 |
|
| 20 |
+
<!-- consumptionqa-examples:start -->
|
| 21 |
+
## Test-set QA examples
|
| 22 |
+
|
| 23 |
+
These four QA excerpts come from the existing held-out consumer financial-choice test set. They retain the original English questions, candidate menus and recorded reference answers. Full household states and original identifiers are omitted. The answers belong to the households in the original records; provide a household state when running inference.
|
| 24 |
+
|
| 25 |
+
[Download the examples as JSON](qa_examples.json).
|
| 26 |
+
|
| 27 |
+
### 1. Active credit-card use
|
| 28 |
+
|
| 29 |
+
Does your household use credit cards? Exclude cards that have not been activated.
|
| 30 |
+
|
| 31 |
+
- `1` — Yes
|
| 32 |
+
- `2` — No
|
| 33 |
+
|
| 34 |
+
**Recorded reference answer:** `2` — No
|
| 35 |
+
|
| 36 |
+
### 2. Demand-deposit account
|
| 37 |
+
|
| 38 |
+
Does your household currently have a renminbi demand-deposit account?
|
| 39 |
+
|
| 40 |
+
- `1` — Yes
|
| 41 |
+
- `2` — No
|
| 42 |
+
|
| 43 |
+
**Recorded reference answer:** `1` — Yes
|
| 44 |
+
|
| 45 |
+
### 3. Stock trading account
|
| 46 |
+
|
| 47 |
+
Does your household currently have a stock trading account?
|
| 48 |
+
|
| 49 |
+
- `1` — Yes
|
| 50 |
+
- `2` — No
|
| 51 |
+
|
| 52 |
+
**Recorded reference answer:** `2` — No
|
| 53 |
+
|
| 54 |
+
### 4. Unmatured time deposits
|
| 55 |
+
|
| 56 |
+
Does your household currently have any renminbi time deposits that have not yet matured?
|
| 57 |
+
|
| 58 |
+
- `1` — Yes
|
| 59 |
+
- `2` — No
|
| 60 |
+
|
| 61 |
+
**Recorded reference answer:** `2` — No
|
| 62 |
+
|
| 63 |
+
<!-- consumptionqa-examples:end -->
|
| 64 |
+
|
| 65 |
## Project background
|
| 66 |
|
| 67 |
S²-MAS is a society-scale multi-agent simulation framework for studying how heterogeneous household decisions become population outcomes under changing conditions. It connects a constructed resource of **500 million household personas**, reusable scenarios, trainable decision models and stateful world updates. FinL-bilingual supplies a reusable decision policy: it scores the complete alternatives for a household state and question, allowing the same policy interface to serve different tasks.
|
Model_Card_ZH.md
CHANGED
|
@@ -17,6 +17,51 @@ tags:
|
|
| 17 |
|
| 18 |
**FinL-bilingual** 是采用多语言 encoder、支持中英文的初始监督双语 checkpoint,其 MLX 格式名为 **FinL-bilingual-mlx**。论文的 A100 单请求速度协议及相对 Jev 的 22.72 倍共享输入 token 吞吐量结果属于这组双语 checkpoint。英语基座完整六轮主要版本为 [FinL](https://huggingface.co/ya9919/FinL) 与 [FinL-mlx](https://huggingface.co/ya9919/FinL-mlx)。
|
| 19 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 20 |
## 项目背景
|
| 21 |
|
| 22 |
S²-MAS 是面向社会规模的多智能体仿真框架,用于研究异质家庭在条件变化下的决策如何汇总为人口层面的结果。它将**5 亿条构造家庭 persona 资源**、可复用场景、可训练决策模型与持续更新的世界状态连接起来。FinL-bilingual 在其中提供可复用的决策策略:根据家庭状态、问题和完整候选选项进行评分,使同一决策接口能够服务不同任务。
|
|
|
|
| 17 |
|
| 18 |
**FinL-bilingual** 是采用多语言 encoder、支持中英文的初始监督双语 checkpoint,其 MLX 格式名为 **FinL-bilingual-mlx**。论文的 A100 单请求速度协议及相对 Jev 的 22.72 倍共享输入 token 吞吐量结果属于这组双语 checkpoint。英语基座完整六轮主要版本为 [FinL](https://huggingface.co/ya9919/FinL) 与 [FinL-mlx](https://huggingface.co/ya9919/FinL-mlx)。
|
| 19 |
|
| 20 |
+
<!-- consumptionqa-examples:start -->
|
| 21 |
+
## 测试集QA示例
|
| 22 |
+
|
| 23 |
+
以下4条QA摘录取自已有消费决策测试集,保留原始英文题干、选项和记录中的参考答案。完整家庭画像及原始标识不公开。答案对应原测试记录中的家庭,示例用于展示任务与标签格式;重新推理时需提供家庭状态。
|
| 24 |
+
|
| 25 |
+
[下载JSON示例](qa_examples.json)。
|
| 26 |
+
|
| 27 |
+
### 1. 已激活信用卡使用
|
| 28 |
+
|
| 29 |
+
Does your household use credit cards? Exclude cards that have not been activated.
|
| 30 |
+
|
| 31 |
+
- `1` — Yes
|
| 32 |
+
- `2` — No
|
| 33 |
+
|
| 34 |
+
**参考答案:** `2` — No
|
| 35 |
+
|
| 36 |
+
### 2. 活期存款账户
|
| 37 |
+
|
| 38 |
+
Does your household currently have a renminbi demand-deposit account?
|
| 39 |
+
|
| 40 |
+
- `1` — Yes
|
| 41 |
+
- `2` — No
|
| 42 |
+
|
| 43 |
+
**参考答案:** `1` — Yes
|
| 44 |
+
|
| 45 |
+
### 3. 股票交易账户
|
| 46 |
+
|
| 47 |
+
Does your household currently have a stock trading account?
|
| 48 |
+
|
| 49 |
+
- `1` — Yes
|
| 50 |
+
- `2` — No
|
| 51 |
+
|
| 52 |
+
**参考答案:** `2` — No
|
| 53 |
+
|
| 54 |
+
### 4. 未到期定期存款
|
| 55 |
+
|
| 56 |
+
Does your household currently have any renminbi time deposits that have not yet matured?
|
| 57 |
+
|
| 58 |
+
- `1` — Yes
|
| 59 |
+
- `2` — No
|
| 60 |
+
|
| 61 |
+
**参考答案:** `2` — No
|
| 62 |
+
|
| 63 |
+
<!-- consumptionqa-examples:end -->
|
| 64 |
+
|
| 65 |
## 项目背景
|
| 66 |
|
| 67 |
S²-MAS 是面向社会规模的多智能体仿真框架,用于研究异质家庭在条件变化下的决策如何汇总为人口层面的结果。它将**5 亿条构造家庭 persona 资源**、可复用场景、可训练决策模型与持续更新的世界状态连接起来。FinL-bilingual 在其中提供可复用的决策策略:根据家庭状态、问题和完整候选选项进行评分,使同一决策接口能够服务不同任务。
|
README.md
CHANGED
|
@@ -17,6 +17,51 @@ tags:
|
|
| 17 |
|
| 18 |
**FinL-bilingual** is the bilingual initial-supervision checkpoint, with a multilingual encoder and English/Chinese support. Its MLX runtime is **FinL-bilingual-mlx**. The paper's A100 single-request timing protocol and 22.72× Jev shared-input-token-throughput result belong to this bilingual checkpoint pair. The main English-base six-round checkpoints are [FinL](https://huggingface.co/ya9919/FinL) and [FinL-mlx](https://huggingface.co/ya9919/FinL-mlx).
|
| 19 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 20 |
## Project background
|
| 21 |
|
| 22 |
S²-MAS is a society-scale multi-agent simulation framework for studying how heterogeneous household decisions become population outcomes under changing conditions. It connects a constructed resource of **500 million household personas**, reusable scenarios, trainable decision models and stateful world updates. FinL-bilingual supplies a reusable decision policy: it scores the complete alternatives for a household state and question, allowing the same policy interface to serve different tasks.
|
|
|
|
| 17 |
|
| 18 |
**FinL-bilingual** is the bilingual initial-supervision checkpoint, with a multilingual encoder and English/Chinese support. Its MLX runtime is **FinL-bilingual-mlx**. The paper's A100 single-request timing protocol and 22.72× Jev shared-input-token-throughput result belong to this bilingual checkpoint pair. The main English-base six-round checkpoints are [FinL](https://huggingface.co/ya9919/FinL) and [FinL-mlx](https://huggingface.co/ya9919/FinL-mlx).
|
| 19 |
|
| 20 |
+
<!-- consumptionqa-examples:start -->
|
| 21 |
+
## Test-set QA examples
|
| 22 |
+
|
| 23 |
+
These four QA excerpts come from the existing held-out consumer financial-choice test set. They retain the original English questions, candidate menus and recorded reference answers. Full household states and original identifiers are omitted. The answers belong to the households in the original records; provide a household state when running inference.
|
| 24 |
+
|
| 25 |
+
[Download the examples as JSON](qa_examples.json).
|
| 26 |
+
|
| 27 |
+
### 1. Active credit-card use
|
| 28 |
+
|
| 29 |
+
Does your household use credit cards? Exclude cards that have not been activated.
|
| 30 |
+
|
| 31 |
+
- `1` — Yes
|
| 32 |
+
- `2` — No
|
| 33 |
+
|
| 34 |
+
**Recorded reference answer:** `2` — No
|
| 35 |
+
|
| 36 |
+
### 2. Demand-deposit account
|
| 37 |
+
|
| 38 |
+
Does your household currently have a renminbi demand-deposit account?
|
| 39 |
+
|
| 40 |
+
- `1` — Yes
|
| 41 |
+
- `2` — No
|
| 42 |
+
|
| 43 |
+
**Recorded reference answer:** `1` — Yes
|
| 44 |
+
|
| 45 |
+
### 3. Stock trading account
|
| 46 |
+
|
| 47 |
+
Does your household currently have a stock trading account?
|
| 48 |
+
|
| 49 |
+
- `1` — Yes
|
| 50 |
+
- `2` — No
|
| 51 |
+
|
| 52 |
+
**Recorded reference answer:** `2` — No
|
| 53 |
+
|
| 54 |
+
### 4. Unmatured time deposits
|
| 55 |
+
|
| 56 |
+
Does your household currently have any renminbi time deposits that have not yet matured?
|
| 57 |
+
|
| 58 |
+
- `1` — Yes
|
| 59 |
+
- `2` — No
|
| 60 |
+
|
| 61 |
+
**Recorded reference answer:** `2` — No
|
| 62 |
+
|
| 63 |
+
<!-- consumptionqa-examples:end -->
|
| 64 |
+
|
| 65 |
## Project background
|
| 66 |
|
| 67 |
S²-MAS is a society-scale multi-agent simulation framework for studying how heterogeneous household decisions become population outcomes under changing conditions. It connects a constructed resource of **500 million household personas**, reusable scenarios, trainable decision models and stateful world updates. FinL-bilingual supplies a reusable decision policy: it scores the complete alternatives for a household state and question, allowing the same policy interface to serve different tasks.
|
qa_examples.json
ADDED
|
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"title": "Selected held-out ConsumptionQA excerpts",
|
| 3 |
+
"content": "Question, option menu and recorded reference answer. Household states and original identifiers are omitted. These are QA excerpts, not complete inference requests.",
|
| 4 |
+
"examples": [
|
| 5 |
+
{
|
| 6 |
+
"example_id": "example-01",
|
| 7 |
+
"topic": "Active credit-card use",
|
| 8 |
+
"split": "test",
|
| 9 |
+
"task": "ConsumptionQA",
|
| 10 |
+
"question_type": "choice",
|
| 11 |
+
"question": "Does your household use credit cards? Exclude cards that have not been activated.",
|
| 12 |
+
"options": {
|
| 13 |
+
"1": "Yes",
|
| 14 |
+
"2": "No"
|
| 15 |
+
},
|
| 16 |
+
"reference_choice": "2",
|
| 17 |
+
"reference_answer": "No"
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"example_id": "example-02",
|
| 21 |
+
"topic": "Demand-deposit account",
|
| 22 |
+
"split": "test",
|
| 23 |
+
"task": "ConsumptionQA",
|
| 24 |
+
"question_type": "choice",
|
| 25 |
+
"question": "Does your household currently have a renminbi demand-deposit account?",
|
| 26 |
+
"options": {
|
| 27 |
+
"1": "Yes",
|
| 28 |
+
"2": "No"
|
| 29 |
+
},
|
| 30 |
+
"reference_choice": "1",
|
| 31 |
+
"reference_answer": "Yes"
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"example_id": "example-03",
|
| 35 |
+
"topic": "Stock trading account",
|
| 36 |
+
"split": "test",
|
| 37 |
+
"task": "ConsumptionQA",
|
| 38 |
+
"question_type": "choice",
|
| 39 |
+
"question": "Does your household currently have a stock trading account?",
|
| 40 |
+
"options": {
|
| 41 |
+
"1": "Yes",
|
| 42 |
+
"2": "No"
|
| 43 |
+
},
|
| 44 |
+
"reference_choice": "2",
|
| 45 |
+
"reference_answer": "No"
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"example_id": "example-04",
|
| 49 |
+
"topic": "Unmatured time deposits",
|
| 50 |
+
"split": "test",
|
| 51 |
+
"task": "ConsumptionQA",
|
| 52 |
+
"question_type": "choice",
|
| 53 |
+
"question": "Does your household currently have any renminbi time deposits that have not yet matured?",
|
| 54 |
+
"options": {
|
| 55 |
+
"1": "Yes",
|
| 56 |
+
"2": "No"
|
| 57 |
+
},
|
| 58 |
+
"reference_choice": "2",
|
| 59 |
+
"reference_answer": "No"
|
| 60 |
+
}
|
| 61 |
+
]
|
| 62 |
+
}
|