Configuring Detection Thresholds

Detection thresholds set the confidence level at which Discuse flags content, balancing false positives against false negatives. In Discuse these thresholds are project settings configured in the dashboard (or the settings API), and the API request only toggles which checks run — it does not carry numeric thresholds. This guide explains how the trade-off works and how to choose values for your platform.

How Discuse thresholds work

Each check returns a confidence score from 0.0 to 1.0 and a hit flag. The hit is set when the score meets or exceeds the threshold you configured for that category in your project. The configurable thresholds are:

Threshold (project setting)	Applies to
threshold_sentiment	Overall negative-sentiment cut-off
threshold_toxicity	Toxic text
threshold_profanity	Profanity
threshold_threat	Threats
threshold_insult	Insults
threshold_spam	Spam classifier confidence
threshold_images	Explicit imagery (overall)
threshold_images_porn	Pornographic imagery
threshold_images_sexual	Sexually suggestive imagery

These are the names exposed by the project settings API; the per-request settings object only contains on/off check_* toggles plus expected_language. You change thresholds in the dashboard, not per request.

The trade-off

Lower Threshold = More Strict
├── More content flagged
├── Higher false positive rate
├── Fewer harmful posts slip through
└── More user friction

Higher Threshold = More Permissive
├── Less content flagged
├── Lower false positive rate
├── More harmful posts may slip through
└── Better user experience

Visualizing the trade-off

                False Positives ─────────────────────────►
                     Few                              Many
               ┌─────────────────────────────────────────┐
False      Few │   ◄─── Ideal Zone                      │
Negatives      │        (High threshold,                │
               │         low false rates)               │
      │        │                                        │
      │        │              Your platform's           │
      │        │              optimal point →  ●        │
      ▼        │                                        │
          Many │                      Too permissive ──►│
               └─────────────────────────────────────────┘
                       Threshold: 0.3   0.5   0.7   0.9

What thresholds suit which platform?

The values below are starting points expressed in Discuse's threshold names (threshold_toxicity, threshold_images_porn, and so on). Treat them as a baseline to tune against your own false-positive data, not as guaranteed-correct numbers. A lower threshold flags more aggressively.

Social media platforms

General-purpose social platforms need balanced moderation:

const SOCIAL_MEDIA_THRESHOLDS = {
  threshold_toxicity: 0.7,
  threshold_profanity: 0.6,
  threshold_threat: 0.5,        // Lower (stricter) for threats
  threshold_insult: 0.7,
  threshold_spam: 0.75,
  threshold_images_porn: 0.6,
  threshold_images_sexual: 0.8  // More permissive for suggestive content
};

Professional / business platforms

Business contexts usually want stricter moderation:

const PROFESSIONAL_THRESHOLDS = {
  threshold_toxicity: 0.5,
  threshold_profanity: 0.4,
  threshold_threat: 0.3,
  threshold_insult: 0.5,
  threshold_spam: 0.6,
  threshold_images_porn: 0.3,
  threshold_images_sexual: 0.5
};

Gaming communities

Gaming platforms may tolerate more banter while staying strict on real threats:

const GAMING_THRESHOLDS = {
  threshold_toxicity: 0.8,
  threshold_profanity: 0.85,    // Banter allowed
  threshold_threat: 0.5,        // Still strict on real threats
  threshold_insult: 0.8,
  threshold_spam: 0.8,
  threshold_images_porn: 0.6,
  threshold_images_sexual: 0.9
};

Children's platforms

Platforms for minors call for the strictest settings:

const CHILDRENS_THRESHOLDS = {
  threshold_toxicity: 0.3,
  threshold_profanity: 0.2,
  threshold_threat: 0.2,
  threshold_insult: 0.3,
  threshold_spam: 0.5,
  threshold_images_porn: 0.1,   // Maximum strictness
  threshold_images_sexual: 0.2
};

Can I vary thresholds dynamically?

Discuse stores one set of thresholds per project, so per-user or per-context variation lives in your application. The patterns below compute an effective threshold on your side; you then either route to different projects (each with its own configured thresholds) or apply the comparison yourself against the scores Discuse returns.

User trust levels

Adjust effective thresholds based on user reputation:

function getThresholds(user) {
  const baseThresholds = PLATFORM_THRESHOLDS;

  const trustMultipliers = {
    new_user: 0.8,       // Stricter (lower effective threshold)
    basic_user: 1.0,     // Standard
    verified_user: 1.15, // Slightly more permissive
    trusted_user: 1.3,   // More permissive
    moderator: 1.5       // Most permissive
  };

  const multiplier = trustMultipliers[user.trustLevel] || 1.0;

  return Object.fromEntries(
    Object.entries(baseThresholds).map(([key, value]) => [
      key,
      typeof value === 'number'
        ? Math.min(value * multiplier, 0.95)
        : adjustNestedThresholds(value, multiplier)
    ])
  );
}

Context-based thresholds

Different content types may need different thresholds:

const CONTEXT_THRESHOLDS = {
  // Public posts visible to everyone
  public_post: {
    toxic: 0.6,
    profanity: 0.5
  },

  // Direct messages between users
  direct_message: {
    toxic: 0.7,      // Slightly more permissive
    profanity: 0.6
  },

  // Comments on public content
  comment: {
    toxic: 0.55,     // Stricter than posts
    profanity: 0.5
  },

  // Profile information
  profile: {
    toxic: 0.5,      // Strict for public-facing content
    profanity: 0.4
  }
};

function getContextThresholds(contentType) {
  return CONTEXT_THRESHOLDS[contentType] || CONTEXT_THRESHOLDS.public_post;
}

Time-based adjustments

Adjust thresholds during high-risk periods:

function getTimeAdjustedThresholds(baseThresholds) {
  const hour = new Date().getHours();
  const dayOfWeek = new Date().getDay();

  // Stricter during off-hours when fewer moderators available
  const isOffHours = hour < 6 || hour > 22;
  const isWeekend = dayOfWeek === 0 || dayOfWeek === 6;

  let multiplier = 1.0;

  if (isOffHours) multiplier *= 0.85;
  if (isWeekend) multiplier *= 0.9;

  return adjustThresholds(baseThresholds, multiplier);
}

Implementing threshold configuration

Centralized configuration

// config/moderation.js — mirrors your Discuse project thresholds so app-side
// routing stays in sync with the values configured in the dashboard.
export const ModerationConfig = {
  thresholds: {
    threshold_toxicity:      parseFloat(process.env.THRESHOLD_TOXICITY || '0.7'),
    threshold_profanity:     parseFloat(process.env.THRESHOLD_PROFANITY || '0.6'),
    threshold_threat:        parseFloat(process.env.THRESHOLD_THREAT || '0.5'),
    threshold_insult:        parseFloat(process.env.THRESHOLD_INSULT || '0.7'),
    threshold_spam:          parseFloat(process.env.THRESHOLD_SPAM || '0.75'),
    threshold_images_porn:   parseFloat(process.env.THRESHOLD_IMAGES_PORN || '0.6'),
    threshold_images_sexual: parseFloat(process.env.THRESHOLD_IMAGES_SEXUAL || '0.8')
  },

  actions: {
    high_confidence: 'auto_block',     // score > 0.95
    medium_confidence: 'human_review', // 0.7 - 0.95
    low_confidence: 'allow_with_flag'  // threshold - 0.7
  }
};

Runtime threshold updates

Allow threshold adjustments without redeployment:

class ModerationService {
  constructor() {
    this.thresholds = defaultThresholds;
    this.loadRemoteConfig();
  }

  async loadRemoteConfig() {
    try {
      const config = await fetch('/api/admin/moderation-config');
      const data = await config.json();
      this.thresholds = data.thresholds;
      console.log('Loaded remote moderation config');
    } catch (error) {
      console.warn('Using default thresholds:', error);
    }
  }

  async checkContent(content, context) {
    const result = await callModerationAPI(content);
    const thresholds = this.getThresholdsForContext(context);

    return this.applyThresholds(result, thresholds);
  }
}

Measuring threshold effectiveness

Key metrics

const MODERATION_METRICS = {
  // Accuracy
  precision: 'True positives / (True positives + False positives)',
  recall: 'True positives / (True positives + False negatives)',
  f1_score: 'Harmonic mean of precision and recall',

  // User impact
  block_rate: 'Content blocked / Total content',
  appeal_rate: 'Appeals filed / Content blocked',
  appeal_success: 'Appeals won / Appeals filed',

  // Operational
  review_queue_size: 'Items waiting for human review',
  review_time: 'Average time to human decision'
};

A/B testing thresholds

Test threshold changes on a subset of traffic:

async function moderateWithExperiment(content, userId) {
  const experiment = getExperiment(userId, 'threshold_test');

  const thresholds = experiment === 'control'
    ? CURRENT_THRESHOLDS
    : EXPERIMENTAL_THRESHOLDS;

  const result = await checkContent(content);
  const decision = applyThresholds(result, thresholds);

  // Log for analysis
  await logExperiment({
    experiment: 'threshold_test',
    variant: experiment,
    content_id: content.id,
    scores: result,
    decision: decision,
    timestamp: Date.now()
  });

  return decision;
}

Analyzing results

-- Calculate precision and recall for each threshold variant
SELECT
  variant,
  COUNT(*) as total_decisions,
  SUM(CASE WHEN blocked AND actually_harmful THEN 1 ELSE 0 END) as true_positives,
  SUM(CASE WHEN blocked AND NOT actually_harmful THEN 1 ELSE 0 END) as false_positives,
  SUM(CASE WHEN NOT blocked AND actually_harmful THEN 1 ELSE 0 END) as false_negatives,
  SUM(CASE WHEN blocked AND actually_harmful THEN 1 ELSE 0 END) * 1.0 /
    NULLIF(SUM(CASE WHEN blocked THEN 1 ELSE 0 END), 0) as precision,
  SUM(CASE WHEN blocked AND actually_harmful THEN 1 ELSE 0 END) * 1.0 /
    NULLIF(SUM(CASE WHEN actually_harmful THEN 1 ELSE 0 END), 0) as recall
FROM moderation_decisions
WHERE experiment = 'threshold_test'
GROUP BY variant;

Threshold tuning workflow

Step 1: Establish a baseline

// Start with conservative thresholds
const INITIAL_THRESHOLDS = {
  threshold_toxicity: 0.5,
  threshold_profanity: 0.5,
  threshold_spam: 0.6
};

Step 2: Collect data

async function logModerationDecision(content, result, decision) {
  await db.insert('moderation_log', {
    content_id: content.id,
    content_hash: hashContent(content.text),
    scores: result.results,
    thresholds_used: currentThresholds,
    decision: decision,
    user_trust_level: content.author.trustLevel,
    created_at: Date.now()
  });
}

Step 3: Analyze false rates

Review blocked content and user appeals to identify:

• False positives: safe content incorrectly blocked
• False negatives: harmful content that wasn't caught

Step 4: Adjust and iterate

// Based on analysis, raise thresholds that fire too often
const ADJUSTED_THRESHOLDS = {
  threshold_toxicity: 0.65,  // Raised after false positives
  threshold_profanity: 0.55,
  threshold_spam: 0.7
};

Step 5: Monitor continuously

Set up alerts for threshold effectiveness:

async function checkModerationHealth() {
  const stats = await getModerationStats(last24Hours);

  // Alert if false positive rate too high
  if (stats.appealSuccessRate > 0.3) {
    alert('High appeal success rate - thresholds may be too strict');
  }

  // Alert if harmful content is getting through
  if (stats.reportedAfterApproval > threshold) {
    alert('Increase in reported content - thresholds may be too permissive');
  }
}

Best practices summary

1. Start conservative: begin with stricter thresholds and loosen them based on data.
2. Use context: different surfaces (public posts, DMs, profiles) warrant different thresholds.
3. Trust levels matter: adjust the effective threshold for user reputation in your app.
4. Measure everything: track precision, recall, and user impact.
5. Iterate continuously: moderation is never "done".
6. Document decisions: keep records of why thresholds changed.
7. Have fallbacks: route borderline cases to human review.

Remember: in Discuse these thresholds are project settings. Change them in the dashboard or via the settings API; the per-request settings object only toggles which check_* run.

Next steps

• AI Content Moderation Guide - Understanding AI moderation
• Scaling Content Moderation - High-volume implementation
• Text Analysis - Text-specific moderation details