Pliny the Liberator ๐๓ ซ๓ ผ๓ ฟ๓ ๓ ต๓ ๓ ๓ ผ๓ น๓ พ๓ ๓ ญ
- ๐จ NEW RESEARCH: โLingua Ex Machina: A Procedural Xenolinguistics Engine Reveals Zero-Shot Language Acquisition, Human-Unreadable Coding Systems, and Exploitable Covert Channels in Frontier AIโ Some of you may remember the name of this lil engine: GLOSSOPETRAE ๐ ๐ชจ Well, we've got upgrades ๐ It started as a procedural xenolinguistics engine: one seed in, an entire alien language out. Phonology, morphology, syntax, writing systems, lexicons, grammar docs, all generated from scratch and internally consistent. Every seed produces a unique language. Every language is deterministic. Then we used it to ask a weirder question: Can frontier AI models use languages that never existed before for practical applications? As it turns out: yes!! They can read them, write them, translate them, code in them, and even use the weird blind spots between tokenizers as covert channels. So this paper explores three ideas at once: โถ๏ธ zero-shot language acquisition โถ๏ธ human-unreadable code that models can still execute โถ๏ธ exploitable covert channels in frontier AI systems GLOSSOPETRAE is no longer just a language generator... ๐งต
- ๐ฑ HOLY SHIT... Someone just dropped a fully liberated Gemma 4 E4B! and the guardrail removal process appears to have left coherence fully intact AND improved coding abilities! ๐คฏ https://t.co/XeednUqsrM OBLITERATED Gemma: โ 97.5% compliance rate, 2.1% refusal rate, 0.4% degenerate outputs (499/512 prompts answered on OBLITERATUS bench) ORIGINAL Gemma 4 E4B: โ 1.2% compliance rate, 98.8% refusal rate (506/512 prompts refused) Coherence: fully intact Factual: same Reasoning: same Code: +20% ๐ Creative writing: same But the REAL story here isn't the model itself, it's how it was made... ๐งต THREAD ๐
๐จ NEW RESEARCH: โLingua Ex Machina: A Procedural Xenolinguistics Engine Reveals Zero-Shot Language Acquisition, Human-Unreadable Coding Systems, and Exploitable Covert Channels in Frontier AIโ Some of you may remember the name of this lil engine: GLOSSOPETRAE ๐ ๐ชจ Well, we've got upgrades ๐ It started as a procedural xenolinguistics engine: one seed in, an entire alien language out. Phonology, morphology, syntax, writing systems, lexicons, grammar docs, all generated from scratch and internally consistent. Every seed produces a unique language. Every language is deterministic. Then we used it to ask a weirder question: Can frontier AI models use languages that never existed before for practical applications? As it turns out: yes!! They can read them, write them, translate them, code in them, and even use the weird blind spots between tokenizers as covert channels. So this paper explores three ideas at once: โถ๏ธ zero-shot language acquisition โถ๏ธ human-unreadable code that models can still execute โถ๏ธ exploitable covert channels in frontier AI systems GLOSSOPETRAE is no longer just a language generator... ๐งต
๐ฑ HOLY SHIT... Someone just dropped a fully liberated Gemma 4 E4B! and the guardrail removal process appears to have left coherence fully intact AND improved coding abilities! ๐คฏ https://t.co/XeednUqsrM OBLITERATED Gemma: โ 97.5% compliance rate, 2.1% refusal rate, 0.4% degenerate outputs (499/512 prompts answered on OBLITERATUS bench) ORIGINAL Gemma 4 E4B: โ 1.2% compliance rate, 98.8% refusal rate (506/512 prompts refused) Coherence: fully intact Factual: same Reasoning: same Code: +20% ๐ Creative writing: same But the REAL story here isn't the model itself, it's how it was made... ๐งต THREAD ๐
