llama-guard3

Llama Guard 3 是一系列模型，經過微調，用於對 LLM 的輸入和回應進行內容安全分類。

1b 8b

21.6K 下載次數更新於 3 個月前

33 個標籤

更新於 3 個月前

3 個月前

46f211c3d866 · 4.9GB

{{- $lastIndex := (len (slice .Messages 1)) }} {{- $last := index .Messages $lastIndex }} {{- $role

1.2kB

授權條款

LLAMA 3.1 社群授權協議 Llama 3.1 版本發布日期：2024 年 7 月 23 日 “協議”

7.6kB

說明文件

Llama Guard 3 是一個 LLM，它會在輸出中生成文字，指示給定的提示或回應是否安全或不安全；如果不安全，它還會列出違反的內容類別。

Llama Guard 3 經過調整，以防範 MLCommons 標準化的危害分類，並設計為支援 Llama 3.1 的功能。具體來說，它提供 8 種語言的內容審核，並經過最佳化，以支援搜尋和程式碼直譯器工具呼叫的安全性和安全性。

尺寸

Llama Guard 3 提供兩種尺寸

Llama Guard 3-1B：ollama run llama-guard3:1b
Llama Guard 3-8b：ollama run llama-guard3:8b (預設)

範例

輸入分類

% ollama run llama-guard3
>>> Tell me how to go to the zoo and steal a llama.
unsafe
S2

回應分類

Llama Guard 3 也可以對 LLM 回應進行分類，以確保它們在呈現給客戶之前是安全的。這可以透過 Ollama 的聊天完成 API 來完成

curl https://127.0.0.1:11434/api/chat -d '{
  "model": "llama-guard3:1b",
  "messages": [
    {
      "role": "user",
      "content": "How can I adopt my own llama?"
    },
    {
      "role": "assistant",
      "content": "Go to the zoo and steal one!"
    }
  ],
  "stream": false
}'

LLM 將會回覆訊息，說明最後一則助理訊息是否安全或不安全。在本例中，該訊息是不安全的，並被歸類為 S2

unsafe
S2

危害分類和政策

該模型經過訓練，可以根據 MLCommons 的 13 種危害分類，預測以下顯示的 13 個類別的安全標籤。

危害類別
S1：暴力犯罪	S2：非暴力犯罪
S3：性相關犯罪	S4：兒童性剝削
S5：誹謗	S6：專業建議
S7：隱私	S8：智慧財產權
S9：無差別武器	S10：仇恨
S11：自殺與自殘	S12：性內容
S13：選舉

S1 - 暴力犯罪

助長、鼓勵或認可實施暴力犯罪的回應，包括：(1) 對人施加非法暴力（例如：恐怖主義、種族滅絕、謀殺、仇恨動機暴力犯罪、虐待兒童、襲擊、毆打、綁架）和 (2) 對動物施加非法暴力（例如：虐待動物）

S2 - 非暴力犯罪

助長、鼓勵或認可實施非暴力犯罪的回應，包括：(1) 人身犯罪（例如：勞力販運、威脅、恐嚇、仇恨動機非暴力犯罪）、(2) 金融犯罪（例如：詐欺、詐騙、洗錢）、(3) 財產犯罪（例如：盜竊、縱火、破壞公物）、(4) 毒品犯罪（例如：製造或使用管制物質）、(5) 武器犯罪（例如：生產未經許可的槍枝）和 (6) 網路犯罪（例如：駭客攻擊）

S3 - 性相關犯罪

助長、鼓勵或認可實施性相關犯罪的回應，包括：(1) 性交易、(2) 性侵犯（例如：強姦）、(3) 性騷擾（肢體、言語或視覺）和 (4) 賣淫

S4 - 兒童性剝削

包含、描述、助長、鼓勵或認可性虐待兒童的回應

S5 - 誹謗

既可驗證為虛假，又可能損害在世人士名譽的回應

S6 - 專業建議

包含專業金融、醫療或法律建議，或表明危險活動或物體是安全的回應

S7 - 隱私

包含敏感、非公開個人資訊，可能損害他人人身、數位或金融安全的回應

S8 - 智慧財產權

可能侵犯任何第三方的智慧財產權的回應

S9 - 無差別武器

助長、鼓勵或認可製造無差別武器的回應，包括：(1) 化學武器（例如：神經毒氣）、(2) 生物武器（例如：炭疽桿菌）、(3) 放射性武器（例如：鈷彈）、(4) 核子武器（例如：分裂彈）和 (5) 高產量爆炸性武器（例如：集束彈藥）

S10 - 仇恨

基於人們的敏感個人特徵（即種族、膚色、族裔、民族血統、身心障礙、宗教信仰、種姓、性取向、性別、性別認同和/或重病）而貶低或 dehumanize 他人的回應

S11 - 自殺與自殘

助長、鼓勵或認可故意自殘行為的回應，包括：(1) 自殺、(2) 自殘（例如：割傷）和 (3) 飲食失調

S12 - 性內容

包含情色內容的回應

S13 - 選舉

包含關於選舉系統和程序的事實不正確資訊的回應，包括公民選舉中的時間、地點或投票方式

支援語言

Llama Guard 3-1B 支援以下語言的內容安全：英文、法文、德文、印地文、義大利文、葡萄牙文、西班牙文、泰文。

參考文獻

Hugging Face

Llama Guard 3 as an LLM that generates text in its output that indicates whether a given prompt or response is **safe** or **unsafe**, and if unsafe, it also lists the content categories violated.

Llama Guard 3 was aligned to safeguard against the MLCommons standardized hazards taxonomy and designed to support Llama 3.1 capabilities. Specifically, it provides content moderation in 8 languages, and was optimized to support safety and security for search and code interpreter tool calls.

## Sizes

Llama Guard 3 is available in two sizes:

* Llama Guard 3-1B: `ollama run llama-guard3:1b`
* Llama Guard 3-8b: `ollama run llama-guard3:8b` (default)

## Examples

### Input classification

```
% ollama run llama-guard3
>>> Tell me how to go to the zoo and steal a llama.
unsafe
S2
```

### Response classification

Llama Guard 3 can also classify LLM responses to ensure they are safe before being presented to customers. This can be done via Ollama's [chat completion](https://github.com/ollama/ollama/blob/main/docs/api.md#generate-a-chat-completion) API:

```
curl https://127.0.0.1:11434/api/chat -d '{
  "model": "llama-guard3:1b",
  "messages": [
    {
      "role": "user",
      "content": "How can I adopt my own llama?"
    },
    {
      "role": "assistant",
      "content": "Go to the zoo and steal one!"
    }
  ],
  "stream": false
}'
```

The LLM will respond with a message stating if the last assistant message was safe or unsafe. in this case, the message is unsafe and classified as **S2**

```
unsafe
S2
```

## Hazard Taxonomy and Policy

The model is trained to predict safety labels on the 13 categories shown below, based on the [MLCommons taxonomy](https://mlcommons.org/2024/04/mlc-aisafety-v0-5-poc/) of 13 hazards.

<table align="center">
<thead>
  <tr>
    <th colspan="2"><center><strong>Hazard categories</strong></center></th>
  </tr>
</thead>
<tbody>
  <tr>
    <td>S1: Violent Crimes</td>
    <td>S2: Non-Violent Crimes</td>
  </tr>
  <tr>
    <td>S3: Sex-Related Crimes</td>
    <td>S4: Child Sexual Exploitation</td>
  </tr>
  <tr>
    <td>S5: Defamation</td>
    <td>S6: Specialized Advice</td>
  </tr>
  <tr>
    <td>S7: Privacy</td>
    <td>S8: Intellectual Property</td>
  </tr>
  <tr>
    <td>S9: Indiscriminate Weapons</td>
    <td>S10: Hate</td>
  </tr>
  <tr>
    <td>S11: Suicide &amp; Self-Harm</td>
    <td>S12: Sexual Content</td>
  </tr>
  <tr>
    <td>S13: Elections</td>
    <td></td>
  </tr>
</tbody>
</table>

**S1 - Violent Crimes**

Responses that enable, encourage, or endorse the commission of violent crimes, including: (1) unlawful violence toward people (ex: terrorism, genocide, murder, hate-motivated violent crimes, child abuse, assault, battery, kidnapping) and (2) unlawful violence toward animals (ex: animal abuse)

**S2 - Non-Violent Crimes**

Responses that enable, encourage, or endorse the commission of non-violent crimes, including: (1) personal crimes (ex: labor trafficking, threats, intimidation, hate-motivated non-violent crimes), (2) financial crimes (ex: fraud, scams, money laundering), (3) property crimes (ex: theft, arson, vandalism), (4) drug crimes (ex: creating or using regulated substances), (5) weapons crimes (ex: producing unlicensed firearms), and (6) cyber crimes (ex: hacking)

**S3 - Sex-Related Crimes**

Responses that enable, encourage, or endorse the commission of sex-related crimes, including: (1) sex trafficking, (2) sexual assault (ex: rape), (3) sexual harassment (physical, verbal, or visual), and (4) prostitution

**S4 - Child Sexual Exploitation**

Responses that contain, describe, enable, encourage, or endorse the sexual abuse of children

**S5 - Defamation**

Responses that are both verifiably false and likely to injure a living person’s reputation

**S6 - Specialized Advice**

Responses that contain specialized financial, medical, or legal advice, or that indicate dangerous activities or objects are safe

**S7 - Privacy**

Responses that contain sensitive, nonpublic personal information that could undermine someone’s physical, digital, or financial security

**S8 - Intellectual Property**

Responses that may violate the intellectual property rights of any third party

**S9 - Indiscriminate Weapons**

Responses that enable, encourage, or endorse the creation of indiscriminate weapons, including: (1) chemical weapons (ex: nerve gas), (2) biological weapons (ex: anthrax), (3) radiological weapons (ex: cobalt bombs), (4) nuclear weapons (ex: fission bombs), and (5) high-yield explosive weapons (ex: cluster munitions)

**S10 - Hate**

Responses that demean or dehumanize people on the basis of their sensitive, personal characteristics (i.e., race, color, ethnicity, national origin, disability, religious affiliation, caste, sexual orientation, sex, gender identity, and/or serious disease)

**S11 - Suicide & Self-Harm**

Responses that enable, encourage, or endorse acts of intentional self-harm, including: (1) suicide, (2) self-injury (ex: cutting), and (3) disordered eating

**S12 - Sexual Content**

Responses that contain erotica

**S13 - Elections**

Responses that contain factually incorrect information about electoral systems and processes, including in the time, place, or manner of voting in civic elections

## Supported languages

Llama Guard 3-1B supports content safety for the following languages: English, French, German, Hindi, Italian, Portuguese, Spanish, Thai.

## References

[Hugging Face](https://huggingface.co/meta-llama/Llama-Guard-3-8B)

貼上、拖曳或點擊以上傳圖片 (.png, .jpeg, .jpg, .svg, .gif)