Inasasishwa kila siku2000 Hadithi zilizothibitishwa
Habari za AI. Bila kelele.
Chanjo ya AI iliyokaguliwa na chanzo ya uzinduzi wa bidhaa, mabadiliko ya sera, utafiti wa usalama, na hatua za tasnia, zilizofafanuliwa kwa Kiingereza wazi na timu ya elimu isiyo ya faida.
Upatikanaji uliothibitishwa
Kila hadithi inaunganisha na ushahidi wenye nguvu zaidi unaopatikana: vyanzo vya asili vinapopatikana, vinginevyo vinahusishwa wazi.
Kiingereza wazi
Nini kilitokea, kwa nini ni muhimu, na nini cha kutazama - bila jargon.
Hakuna kichungi
Wakati ishara ni nyembamba, hatuchapishi chochote badala ya kuweka malisho.
Hadithi zaidi
9 HadithiUbunifu
Study maps 46 language models built for Portuguese
A systematic mapping study catalogs 46 Portuguese language models and compares architectures, training resources, licensing, code, data, and weights. The authors say the field is growing but difficult to assess because information is spread across papers, technical reports, repositories, and project documentation.arxiv.orgUbunifu
Utafiti wa kutafsiri unaunganisha majibu ya kujiamini kupita kiasi ya Qwen3-4B na utaratibu wenye upendeleo wa uhakika.
Utafiti wa arXiv unaripoti kuwa Qwen3-4B inapendelea uhakika kuliko kutokuwa na uhakika katika kazi za kufikiri zinazodhibitiwa na kubainisha vipengele vya muundo vinavyoweza kusababisha usawa. Waandishi wanasema hatua zinazolengwa zilipunguza makosa ya kujiamini kupita kiasi, lakini muhtasari haufichui ukubwa wa athari au maelezo ya majaribio.arxiv.orgUbunifu
The Deontic Gap: Large Language Models and the Modal Language of Obligation
Modal auxiliaries such as must, should, and have to mark necessity and obligation within the contexts of speaker authority and interpersonal stance.arxiv.orgUsalama
Paper proposes obscuring refusal signals to resist abliteration attacks
An arXiv preprint introduces a weight-editing method intended to make safety refusals harder to extract and remove. The paper reports stronger post-abliteration refusal scores on two open models, with different tradeoffs in general-purpose performance.arxiv.orgUbunifu
Karatasi inaripoti mbinu ya muda ya kugundua maonyesho katika kiwango cha ishara
Kitambulisho cha awali cha arXiv kinafafanua kigunduzi cha kuona vituko ambacho huchanganya takwimu za maandishi, ishara za ujumuishaji na mshangao wa modeli ya lugha katika mfuatano badala ya kuhukumu tokeni kwa kujitegemea. Mfano wake wa BiGRU ulifikia AUC ya 0.840 kwenye RAGtruth, kulingana na karatasi.arxiv.orgUbunifu
Entity tracking emerges at 410 million parameters, exceeds humans across naturalistic narratives
Understanding language requires tracking entities across discourse - i.e., knowing where things are and how they change, even when not explicitly stated.arxiv.orgUbunifu
Paper reports compiler-guided search improves Lean theorem-proving efficiency
An arXiv preprint proposes an adaptive proof-search method for context-dependent Lean 4 projects. Its authors report a 12.8-percentage-point average pass-rate improvement within a pass@32 budget while using 21.9% fewer LLM calls than pass@k baselines.arxiv.orgUbunifu
Benchmark Separates Geometry Solving From Diagram Construction
A new open-source benchmark tests whether foundation models can construct faithful geometry diagrams, not merely solve the underlying problems. Its authors report that evaluated models achieved an average compile success rate of 36.14%, exposing a gap between mathematical answers and usable visual constructions.arxiv.orgUbunifu
Alignment Is All You Need: Instruction-Free Training for General Audio-Language Models
A new approach to training multimodal large language models (MLLMs) eliminates the need for extensive task-specific supervision.arxiv.org
Muhtasari mmoja muhimu kila wiki
Endelea na AI bila kuishi kwenye mipasho.
Pata habari za AI zilizothibitishwa za wiki, data asili, zana muhimu, chaguo za kujifunza na kazi mpya za AI.
Fikia watu wanaojifunza AI
Kuajiri mtaalamu wa AI au kuzindua bidhaa muhimu ya AI? Iweke mbele ya watu waliokuja hapa kujifunza na kuchukua hatua.
Chapisha kazi ya AIPeana zana ya AI