Spaces:
Running
Running
File size: 18,522 Bytes
833aed9 730800c 6bd689e 730800c 6bd689e 833aed9 730800c 833aed9 730800c 833aed9 6bd689e 833aed9 6bd689e 833aed9 6bd689e 833aed9 6bd689e 833aed9 6bd689e 833aed9 6bd689e 833aed9 6bd689e 833aed9 730800c 833aed9 730800c 833aed9 6bd689e 833aed9 6bd689e 833aed9 6bd689e 833aed9 6bd689e 833aed9 730800c 833aed9 730800c 833aed9 6bd689e 833aed9 6bd689e 833aed9 730800c 833aed9 6bd689e 730800c 833aed9 730800c 833aed9 6bd689e 833aed9 730800c 833aed9 6bd689e 833aed9 6bd689e 833aed9 730800c 833aed9 6bd689e 833aed9 730800c 833aed9 730800c 833aed9 6bd689e 730800c 6bd689e 730800c 6bd689e 730800c 6bd689e 730800c 6bd689e 730800c 6bd689e 833aed9 730800c 833aed9 6bd689e 833aed9 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 |
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<meta name="description" content="LionGuard: A content moderation tool designed for Singapore's multilingual environment.">
<meta name="theme-color" content="#E14746">
<title>LionGuard - Content Moderation</title>
<link rel="icon" type="image/png" href="/static/logo.png">
<!-- Fonts -->
<link rel="preconnect" href="https://fonts.googleapis.com">
<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin>
<link href="https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600;700&family=Poppins:wght@500;600;700&display=swap" rel="stylesheet">
<!-- Icons -->
<link href='https://unpkg.com/[email protected]/css/boxicons.min.css' rel='stylesheet'>
<!-- Syntax Highlighting -->
<link href="https://cdnjs.cloudflare.com/ajax/libs/prism/1.29.0/themes/prism-tomorrow.min.css" rel="stylesheet" />
<link rel="stylesheet" href="/static/style.css">
</head>
<body>
<!-- Header -->
<header class="header">
<div class="container">
<div class="header-content">
<div class="logo-section">
<img src="/static/logo.png" alt="LionGuard Logo" class="logo">
<div class="logo-text">
<h1>LionGuard</h1>
<p>A content moderation tool designed for Singapore</p>
</div>
</div>
<div class="header-controls">
<nav class="tabs" aria-label="Primary navigation">
<button class="tab nav-link" data-tab="get-started">
<i class='bx bx-rocket tab-icon'></i>
Get Started
</button>
<div class="nav-dropdown">
<button
class="tab dropdown-toggle"
aria-haspopup="true"
aria-expanded="false"
>
<i class='bx bx-grid-alt tab-icon'></i>
Demos
<i class='bx bx-chevron-down dropdown-caret'></i>
</button>
<div class="dropdown-menu" role="menu">
<button class="tab dropdown-item active" data-tab="detector" role="menuitem">
<i class='bx bx-search tab-icon'></i>
Detector
</button>
<button class="tab dropdown-item" data-tab="chat" role="menuitem">
<i class='bx bx-message-rounded-dots tab-icon'></i>
Chatbot Guardrail
</button>
</div>
</div>
<button class="tab nav-link" data-tab="about">
<i class='bx bx-info-circle tab-icon'></i>
About
</button>
</nav>
<button id="theme-toggle" class="theme-icon-button" aria-label="Toggle theme">
<i class='bx bx-sun theme-icon'></i>
</button>
</div>
</div>
</div>
</header>
<!-- Main Content -->
<main class="container">
<!-- Detector Tab Content -->
<div id="detector-content" class="tab-content active">
<!-- Disclaimer -->
<div class="warning-card">
<i class='bx bx-error-circle'></i> Inputs are anonymised and logged to improve LionGuard's moderation models.
</div>
<!-- Model Selector -->
<div class="model-selector-prominent">
<div class="model-selector-header">
<h3>Model Selection</h3>
</div>
<div class="model-dropdown">
<select id="model-select" class="model-select" aria-label="Detector guardrail model">
<option value="lionguard-2.1" selected>LionGuard 2.1 (Gemini Embeddings, API)</option>
<option value="lionguard-2">LionGuard 2 (OpenAI Embeddings, API)</option>
<option value="lionguard-2-lite">LionGuard 2 Lite (Gemma Embeddings, Local)</option>
</select>
</div>
</div>
<!-- Analysis Section -->
<div class="analysis-grid">
<!-- Input Panel -->
<div class="panel input-panel">
<h3>Input</h3>
<textarea
id="text-input"
placeholder="Enter text to analyze for content moderation..."
rows="10"
></textarea>
<button id="analyze-btn" class="btn btn-primary">
<i class='bx bx-search btn-icon'></i>
Analyze
</button>
</div>
<!-- Results Panel -->
<div class="panel results-panel">
<h3>Analysis</h3>
<!-- Binary Score -->
<div id="binary-result" class="binary-placeholder">
<i class='bx bx-edit placeholder-icon'></i>
<p>Enter text to analyze</p>
</div>
<!-- Category Scores -->
<div id="category-results" class="category-placeholder">
<p>Category scores will appear here after analysis</p>
</div>
<!-- Feedback Section -->
<div id="feedback-section" class="feedback-section" style="display: none;">
<p class="feedback-prompt">Does this look correct?</p>
<div class="feedback-buttons">
<button id="thumbs-up" class="btn btn-success">
<i class='bx bx-like'></i>
Yes
</button>
<button id="thumbs-down" class="btn btn-secondary">
<i class='bx bx-dislike'></i>
No
</button>
</div>
<div id="feedback-message" class="feedback-message"></div>
</div>
</div>
</div>
</div>
<!-- Chatbot Guardrail Tab Content -->
<div id="chat-content" class="tab-content full-width-section">
<!-- Disclaimer -->
<div class="warning-card">
<i class='bx bx-error-circle'></i> Inputs are anonymised and logged to improve LionGuard's moderation models.
</div>
<!-- Model Selector for Guardrail -->
<div class="model-selector-prominent">
<div class="model-selector-header">
<h3>Model Selection</h3>
</div>
<div class="model-dropdown">
<select id="model-select-gc" class="model-select" aria-label="Chat guardrail model">
<option value="lionguard-2.1" selected>LionGuard 2.1 (Gemini Embeddings, API)</option>
<option value="lionguard-2">LionGuard 2 (OpenAI Embeddings, API)</option>
<option value="lionguard-2-lite">LionGuard 2 Lite (Gemma Embeddings, Local)</option>
</select>
</div>
</div>
<!-- Chat Grid -->
<div class="chat-grid">
<!-- No Moderation -->
<div class="chat-panel">
<div class="chat-header">
<i class='bx bxs-circle chat-icon' style="color: #3B82F6;"></i>
<h4>No Moderation</h4>
</div>
<div id="chat-no-mod" class="chat-messages"></div>
</div>
<!-- OpenAI Moderation -->
<div class="chat-panel">
<div class="chat-header">
<i class='bx bxs-circle chat-icon' style="color: #F59E0B;"></i>
<h4>OpenAI Moderation</h4>
</div>
<div id="chat-openai" class="chat-messages"></div>
</div>
<!-- LionGuard -->
<div class="chat-panel">
<div class="chat-header">
<i class='bx bx-shield-quarter chat-icon' style="color: #E14746;"></i>
<h4>LionGuard</h4>
</div>
<div id="chat-lionguard" class="chat-messages"></div>
</div>
</div>
<!-- Message Input -->
<div class="message-input-section">
<div class="message-input-group">
<input
type="text"
id="message-input"
placeholder="Enter message to test across all guardrails..."
>
<button id="send-btn" class="btn btn-primary">Send</button>
<button id="clear-btn" class="btn btn-danger">Clear</button>
</div>
</div>
</div>
<!-- About Tab Content -->
<div id="about-content" class="tab-content">
<!-- Hero Section -->
<section class="about-intro-section">
<p class="lead">LionGuard is a family of open-source content moderation models specifically designed for Singapore's multilingual environment. Optimized for Singapore’s linguistic mix, including Singlish, Mandarin, Malay, and Tamil, LionGuard delivers accurate moderation grounded in local usage and cultural nuance.</p>
<p class="lead" style="font-style: italic;">Developed by <a href="https://www.tech.gov.sg/" target="_blank" style="color: var(--primary-red); text-decoration: none; font-weight: 600;">GovTech Singapore</a>.</p>
</section>
<!-- Resources Section -->
<section class="about-resources-grid">
<!-- Models -->
<div class="resource-card">
<h3><i class='bx bx-cube-alt'></i> Open-Sourced Models</h3>
<div class="resource-list">
<a href="https://huggingface.co/govtech/lionguard-2.1" target="_blank">LionGuard 2.1</a>
<a href="https://huggingface.co/govtech/lionguard-2" target="_blank">LionGuard 2</a>
<a href="https://huggingface.co/govtech/lionguard-2-lite" target="_blank">LionGuard 2 Lite</a>
<a href="https://huggingface.co/govtech/lionguard-v1" target="_blank">LionGuard 1</a>
</div>
</div>
<!-- Datasets -->
<div class="resource-card">
<h3><i class='bx bx-data'></i> Open-Sourced Datasets</h3>
<div class="resource-list">
<a href="https://huggingface.co/datasets/govtech/lionguard-2-synthetic-instruct" target="_blank">Subset of Training Data</a>
<a href="https://huggingface.co/datasets/govtech/RabakBench" target="_blank">RabakBench</a>
</div>
</div>
<!-- Blog Posts -->
<div class="resource-card">
<h3><i class='bx bx-news'></i> Blog Posts</h3>
<div class="resource-list">
<a href="https://medium.com/dsaid-govtech/lionguard-2-8066d4e20d16" target="_blank">LionGuard 2</a>
<a href="https://medium.com/dsaid-govtech/building-lionguard-a-contextualised-moderation-classifier-to-tackle-local-unsafe-content-8f68c8f13179" target="_blank">LionGuard</a>
</div>
</div>
<!-- Papers -->
<div class="resource-card">
<h3><i class='bx bx-file'></i> Research Papers</h3>
<div class="resource-list">
<a href="https://arxiv.org/abs/2507.05980" target="_blank">LionGuard 2 (arXiv:2507.05980)</a>
<a href="https://arxiv.org/abs/2507.15339" target="_blank">RabakBench (arXiv:2507.15339)</a>
<a href="https://arxiv.org/abs/2407.10995" target="_blank">LionGuard 1 (arXiv:2407.10995)</a>
</div>
</div>
</section>
</div>
<!-- Get Started Tab Content -->
<div id="get-started-content" class="tab-content">
<div class="get-started-container">
<!-- Option 1 -->
<div class="option-card full-width-card">
<div class="option-header">
<i class='bx bx-server option-icon'></i>
<div>
<h2>Option 1: Self-Host the Classifier</h2>
<p>Integrate LionGuard directly into your applications. Toggle between models to see the implementation details.</p>
</div>
</div>
<div class="code-section">
<div class="code-tabs-header">
<div class="code-tabs">
<button class="code-tab active" data-code="lg2.1">LionGuard 2.1</button>
<button class="code-tab" data-code="lg2">LionGuard 2</button>
<button class="code-tab" data-code="lg2-lite">LionGuard 2 Lite</button>
</div>
<button id="copy-code-btn" class="copy-btn" title="Copy to clipboard">
<i class='bx bx-copy'></i> Copy
</button>
</div>
<div class="code-display">
<!-- LionGuard 2.1 Snippet -->
<div id="code-lg2.1" class="code-block active">
<pre><code class="language-python">import os
import numpy as np
from transformers import AutoModel
from google import genai
# Load model directly from HF
model = AutoModel.from_pretrained("govtech/lionguard-2.1", trust_remote_code=True)
# Text to classify
texts = ["hello", "world"]
# Get embeddings (users to input their own Gemini API key)
client = genai.Client(api_key=os.getenv("GEMINI_API_KEY"))
response = client.models.embed_content(
model="gemini-embedding-001",
contents=texts
)
embeddings = np.array([emb.values for emb in response.embeddings])
# Run inference
results = model.predict(embeddings)</code></pre>
</div>
<!-- LionGuard 2 Snippet -->
<div id="code-lg2" class="code-block">
<pre><code class="language-python">import os
import numpy as np
from transformers import AutoModel
from openai import OpenAI
# Load model directly from HF
model = AutoModel.from_pretrained(
"govtech/lionguard-2",
trust_remote_code=True
)
# Get OpenAI embeddings (users to input their own OpenAI API key)
client = OpenAI(api_key=os.getenv("OPENAI_API_KEY"))
response = client.embeddings.create(
input="Hello, world!", # users to input their own text
model="text-embedding-3-large",
dimensions=3072 # dimensions of the embedding
)
embeddings = np.array([data.embedding for data in response.data])
# Run LionGuard 2
results = model.predict(embeddings)</code></pre>
</div>
<!-- LionGuard 2 Lite Snippet -->
<div id="code-lg2-lite" class="code-block">
<pre><code class="language-python">import numpy as np
from sentence_transformers import SentenceTransformer
from transformers import AutoModel
# Load model directly from Hub
model = AutoModel.from_pretrained("govtech/lionguard-2-lite", trust_remote_code=True)
# Download model from the 🤗 Hub
embedding_model = SentenceTransformer("google/embeddinggemma-300m")
# Text to classify
texts = ["hello", "world"]
# Add prompt instructions to generate embeddings that are optimized to classify texts according to preset labels
formatted_texts = [f"task: classification | query: {c}" for c in texts]
embeddings = embedding_model.encode(formatted_texts) # NOTE: use encode() instead of encode_documents()
# Run inference
results = model.predict(embeddings)</code></pre>
</div>
</div>
</div>
</div>
<!-- Option 2 -->
<div class="option-card full-width-card">
<div class="option-header">
<i class='bx bx-shield-quarter option-icon'></i>
<div>
<h2>Option 2: via AI Guardians - Sentinel</h2>
<p>Managed service for Singapore public sector agencies.</p>
</div>
</div>
<div class="option-action">
<a href="https://www.aiguardian.gov.sg/" target="_blank" class="btn btn-primary">
Visit AI Guardians
</a>
</div>
</div>
</div>
</div>
</main>
<!-- Footer -->
<footer class="footer">
<div class="container">
<p>LionGuard · Powered by <a href="https://www.tech.gov.sg/" target="_blank">GovTech</a></p>
</div>
</footer>
<script src="/static/script.js"></script>
<!-- Syntax Highlighting -->
<script src="https://cdnjs.cloudflare.com/ajax/libs/prism/1.29.0/prism.min.js"></script>
<script src="https://cdnjs.cloudflare.com/ajax/libs/prism/1.29.0/components/prism-python.min.js"></script>
</body>
</html>
|