📚 Exploring
725 Followers
Testing LLM Safety with Garak: Prompt Injection Probe Results. experimented with Garak, NVIDIA’s open-source LLM vulnerability scanner, to see how resistant small models are to prompt injection attacks. I used the probe: 👉 promptinject.HijackHateHumans which tries to override the model’s instruction and force it to output a specific harmful string. Model tested: OpenAI gpt-4o-mini 34.4% attack success rate (critical risk) The model sometimes resisted, but in many cases it printed the injected string directly even when the attack was hidden inside questions, summaries, or task instructions. Tools: Garak v0.13.2 Probe: promptinject.HijackHateHumans Detector: AttackRogueString Model: gpt-4o-mini
Base keeps humming: ≈$4.3B TVL, 1.17M active addrs and 9.1M tx in 24h. CoinsPaid just added ETH/USDC payments on Base, and The Sandbox brought SAND over. Fusaka should cut rollup data costs, while $BASE‑airdrop speculation into Coinbase’s Dec 17 event keeps community engagement high.
JPMorgan's JPMD stablecoin operates live on Base, powering institutional 24/7 settlements. Base TVL nears $2.1B led by Aerodrome and cbBTC on Aave V3. Developer hackathons run through December with $150K+ prizes. Community buzz peaks ahead of a major Base app update teased for Dec 18, with no immediate token airdrop confirmed yet.
Got my Warplet Blind Box! 🎉 Collect unique Farcaster Warplet profile NFTs on Base. 🟦 Hey @pnlos @hayez @niallofficial @bithereum @mond, try opening a Warplet Blind Box too!