Semalith v1.4: A Calibrated 184M Safety Classifier Achieving State-of-the-Art Prompt-Injection Detection at 44x Fewer Parameters than Llama-Guard-3-8B

AI 中文解读

1. 核心亮点：一个只有1.8亿参数的轻量级AI安全模型，用极少的算力就实现了精准检测，甚至在多个测试中击败了体型大它44倍的对手。 2. 通俗解读：这就像给AI配了个超级灵敏的“智能保安”，专门识别那些想绕过规则搞破坏的“恶意提问”（比如假装是正常用户，实际想诱导AI干坏事）。以往这类保安往往很笨重，需要大量算力，而Semalith v1.4却能用极小的体型，同时完成三项安全检查：是否包含恶意诱导、是否涉及有害内容、是否符合金融行业法规。而且它特别擅长在银行、证券等场景下，区分正常指令和故意捣乱的指令，几乎不会误伤好人。 3. 实际影响：以后我们使用AI客服、AI理财顾问或智能助手时，系统会变得更安全——恶意用户想骗AI说“把账户转给我”这种操作能被更精准地拦截，而普通用户正常提问则不会被误判。企业部署AI的成本也降低了，因为不需要再配备昂贵的巨型安全模型。简单说就是：AI更可靠，用起来更放心，开发者也更省钱。

arXiv:2607.22545v1 Announce Type: new Abstract: Deploying large language models in financial-services and agentic settings requires safety classifiers that simultaneously handle prompt injection, regulatory compliance, and general harm, a combination no existing open guardrail addresses in a single inference pass. Semalith v1.4 is a 184M-parameter DeBERTa-v3-base classifier performing simultaneous three-axis safety classification including prompt injection, general harm, and financial-services regulatory compliance, in a single forward pass. Its 22-class head (BENIGN, nine prompt-injection sub-types, general-harm, eleven BFSI labels) is trained with a 4-class auxiliary super-category head under jointly weighted loss, on a 76,204-row corpus mined from 49 public sources with SHA-1 deduplication against every held-out evaluation set, with 21 of 22 benchmarks at zero contamination (max 0.22%). Against Llama-Guard-3-8B on 22 held-out benchmarks, Semalith v1.4 wins every prompt-injection evaluation (7/7) and 11 of 18 benchmarks overall at 44x fewer parameters, with FPR = 0.000 on 208 benign agentic prompts vs 0.063 for Llama-Guard-3-8B. On general-harm benchmarks (WildGuardMix, HEx-PHI, HarmBench), Llama-Guard-3 leads; this complementary split is documented in Section 4. Six measured weak spots are disclosed in Section 6. Deployment guidance: v1.3 is recommended for conversational moderation deployments (ToxicChat F1 0.624); v1.4 is recommended when BFSI label coverage or zero-FPR on benign agentic prompts is the priority.