<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Guardrails &amp; AI Safety — FoundationIndex</title>
    <link>https://foundationindex.ustunozgur.com/s/guardrails</link>
    <description>Runtime safety checks, input/output classifiers, and programmable rules that wrap AI models to block prompt injections, jailbreaks, and policy violations in production.</description>
    <language>en</language>
    <atom:link href="https://foundationindex.ustunozgur.com/s/guardrails/feed.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Guardrails &amp; AI Safety · Paper · 2025-02-03</title>
      <link>https://foundationindex.ustunozgur.com/s/guardrails#guardrails-constitutional-classifiers-2025</link>
      <guid>https://foundationindex.ustunozgur.com/s/guardrails#guardrails-constitutional-classifiers-2025</guid>
      <pubDate>Mon, 03 Feb 2025 00:00:00 GMT</pubDate>
      <description>&lt;p&gt;Anthropic published Constitutional Classifiers, training streaming input/output safeguards on synthetic constitution-guided data to block universal jailbreaks with minimal over-refusal.&lt;/p&gt;&lt;p&gt;Sources: &lt;a href=&quot;https://www.anthropic.com/research/constitutional-classifiers&quot;&gt;Constitutional Classifiers: Defending against universal jailbreaks across thousands of hours of red teaming&lt;/a&gt;&lt;/p&gt;</description>
    </item>
    <item>
      <title>Guardrails &amp; AI Safety · Paper · 2023-12-07</title>
      <link>https://foundationindex.ustunozgur.com/s/guardrails#guardrails-llama-guard-2023</link>
      <guid>https://foundationindex.ustunozgur.com/s/guardrails#guardrails-llama-guard-2023</guid>
      <pubDate>Thu, 07 Dec 2023 00:00:00 GMT</pubDate>
      <description>&lt;p&gt;Inan et al. at Meta released Llama Guard, an open-weight input/output safety classifier fine-tuned from Llama to score prompts and responses against customizable safety taxonomies.&lt;/p&gt;&lt;p&gt;Sources: &lt;a href=&quot;https://arxiv.org/abs/2312.06674&quot;&gt;Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations&lt;/a&gt;&lt;/p&gt;</description>
    </item>
    <item>
      <title>Guardrails &amp; AI Safety · Paper · 2023-10-16</title>
      <link>https://foundationindex.ustunozgur.com/s/guardrails#guardrails-nemo-2023</link>
      <guid>https://foundationindex.ustunozgur.com/s/guardrails#guardrails-nemo-2023</guid>
      <pubDate>Mon, 16 Oct 2023 00:00:00 GMT</pubDate>
      <description>&lt;p&gt;Rebedea et al. at NVIDIA published NeMo Guardrails, an open-source toolkit using Colang programmable rails to enforce topical boundaries, fact-checking, and safe tool execution around LLM apps.&lt;/p&gt;&lt;p&gt;Sources: &lt;a href=&quot;https://arxiv.org/abs/2310.10501&quot;&gt;NeMo Guardrails: A Toolkit for Controllable and Safe LLM Applications with Programmable Rails&lt;/a&gt;&lt;/p&gt;</description>
    </item>
  </channel>
</rss>
