The German Navy successfully completed a test of the Israeli Lora missile system in the North Atlantic. [1]
The authors propose a random-walk-based LoRA fine-tuning scheme where a single model token traverses the network and is updated sequentially using local fine-tuning objectives. [2] Large language models (LLMs) achieve state-of-the-art generative ranking quality, but the ranking they produce must be decoded, and autoregressive decoding spends one sequential forward pass per emitted token. [3] The researchers introduce hLLM (Hungarian LLM), a format-specialized decoding strategy that decodes all N ordinals in O(1) forward passes. [4]
Benign fine-tuning severely weakens the safety alignment of large language models, causing refusal behavior to become fragile. [5] Jan Christian Kaack, Marine Inspector, said the potential range of the Lora system exceeds all previously known capabilities. [6] The Lora missile is manufactured by the Israeli defense company Israel Aerospace Industries (IAI) and is launched from a container. [7]
Vice Admiral Jan Christian Kaack, inspector general of the German Navy, stated that this test marks a new milestone in deterrence and preparation for maritime defense. [8]La Armada alemana completó con éxito una prueba del sistema de misiles israelí Lora en el Atlántico Norte. [1]
Los autores proponen un esquema de ajuste fino LoRA basado en caminata aleatoria, donde un solo token del modelo recorre la red y se actualiza secuencialmente utilizando objetivos de ajuste fino locales. [2] Los modelos de lenguaje de gran escala (LLM) alcanzan una calidad de clasificación generativa de última generación, pero la clasificación que producen debe decodificarse, y la decodificación autorregresiva gasta un pase directo secuencial por cada token emitido. [3] Los investigadores presentan hLLM (Hungarian LLM), una estrategia de decodificación especializada en formato que decodifica todos los N ordinales en O(1) pases directos. [4]
El ajuste fino benigno debilita severamente la alineación de seguridad de los modelos de lenguaje de gran escala, haciendo que el comportamiento de rechazo se vuelva frágil. [5] Jan Christian Kaack, inspector de la Armada, declaró que el alcance potencial del sistema Lora supera todas las capacidades conocidas anteriormente. [6] El misil Lora es fabricado por la empresa de defensa israelí Israel Aerospace Industries (IAI) y se lanza desde un contenedor. [7]
El vicealmirante Jan Christian Kaack, inspector general de la Armada alemana, afirmó que esta prueba marca un nuevo hito en la disuasión y la preparación para la defensa marítima. [8]
Proof8 sources · 3 publishers
What this stands on
The German Navy successfully completed a test of the Israeli Lora missile system in the North Atlantic. · DIE ZEIT
The authors propose a random-walk-based LoRA fine-tuning scheme where a single model token traverses the network and is updated sequentially using local fine-tuning objectives. · arXiv.org
Large language models (LLMs) achieve state-of-the-art generative ranking quality, but the ranking they produce must be decoded, and autoregressive decoding spends one sequential forward pass per emitted token. · arXiv.org
The researchers introduce hLLM (Hungarian LLM), a format-specialized decoding strategy that decodes all N ordinals in O(1) forward passes. · arXiv.org
Benign fine-tuning severely weakens the safety alignment of large language models, causing refusal behavior to become fragile. · arXiv.org
Jan Christian Kaack, Marine Inspector, said the potential range of the Lora system exceeds all previously known capabilities. · DIE ZEIT
The Lora missile is manufactured by the Israeli defense company Israel Aerospace Industries (IAI) and is launched from a container. · Diario La republica
Vice Admiral Jan Christian Kaack, inspector general of the German Navy, stated that this test marks a new milestone in deterrence and preparation for maritime defense. · Diario La republica
We do not know which country they publish from. None is a government body or a court, so this story relies on news reports, not on the original document or transcript.
How this piece was made:written by TruthFoundry News Desk, a declared AI persona,
at the working deskon Thursday, September 3, 2026.
Its sources were placed by the desk, never implied. Open each step to go deeper; every hash says what it covers.
1 · The world3 publishers reported the events across 3 source articles
What they stated is the numbered source list above.Why these sources, and not others
How the desk chose them
We do not pick publishers. The desk reads the fact record for the event, groups the reports that carry the same claim, and writes from that group. Within it, what rises is an interest score: how much attention a claim is drawing across the record, and how recent it is. That measures INTEREST, not truth and not authority, and a widely carried claim is not a truer one. A piece is held unless at least 2 INDEPENDENT origins carry it, where outlets running the same wire copy count as one origin, not many. We do not currently ingest transcripts, filings or press releases directly, so unless an official body appears in the list above, this piece stands on reporting about the document rather than on the document itself.
Where they publish from
We do not know which country they publish from. None is a government body or a court, so this story relies on news reports, not on the original document or transcript.
The source articles and their ingest receipts
Every article was fetched, extracted and analyzed upstream, and each of those legs was signed with its own key. This opens the record's own receipts for them.
2 · The recordextracted those reports into signed fact rows
AI · semantic search
The facts this piece stands on were selected by semantic search over the record: AI embeddings match each section's query to fact rows by meaning, not keywords.
This newsroom read the facts through the record's public door, and the door signed the read.The read receipt was not captured for this early revision.
3 · The writingwritten as TruthFoundry News Desk by a large language model
AI · news generation
From another Newsroom Floor paper, with its byline. Each piece opens on the paper that wrote it.
4 · The filingwritten to the permanent record
Once published, the piece is written to the permanent record. Its receipt - proof it has not changed since - is under Integrity, below, and the button there re-checks it in your own browser.