Assine o Pro, sem anúncios
Faça upgrade para uma leitura sem interrupções e acesso prioritário a novas fontes.
Ver preçosGemini 3.5 Live Translate brings near real-time, natural speech translation to Google AI Studio, Google Translate and Google Meet.
Gemini 3.5 is built to help you execute complex, agentic workflows.
Explore how AlphaEvolve's Gemini-powered algorithms are driving impact across business, infrastructure, and science.
Gemini Robotics ER 1.6: Enhancing spatial reasoning and multi-view understanding for autonomous robotics.
The Gemini app now features our most advanced music generation model Lyria 3, empowering anyone to make 30-second tracks using text or images.
Gemini 3 Flash offers frontier intelligence built for speed at a fraction of the cost.
Gemini 2.5 Flash-Lite, previously in preview, is now stable and generally available. This cost-efficient model provides high quality in a small size, and includes 2.5 family features like a 1 million-token context window and multimodality.
Gemini 2.5 Deep Think achieves breakthrough performance at the world’s most prestigious computer programming competition, demonstrating a profound leap in abstract problem solving.
The third post in our semantic-layer series builds a customer 360 across Amazon Redshift and Amazon Aurora with Strategy Mosaic, exposes it to a Strands agent on Amazon Bedrock AgentCore through the Mosaic MCP server, and enforces per-user row and column security end-to-end.
This paper introduces BanglaBBQ, a question-answering bias benchmark designed to capture stereotypes grounded in the Bangladeshi context, in both Bangla and English. Following KoBBQ and PakBBQ, BanglaBBQ adapts the original Bias Benchmark for Question Answering (BBQ) by classifying its templates into three groups: directly translated (DT), modified to align target groups or contexts with Bangladesh (TM), and removed when deemed unsuitable for the Bangladeshi context (SR). This process retains 140 templates across six categories, comprising 76 DT and 64 TM templates. Five newly authored (NA) categories, based on stereotypes common in Bangladeshi society, are then added: marital status, age and political perception, regional origin, educational background, and mental-health status. Native Bangladeshi annotators judged all 466 newly authored templates, and 372 were retained because a strict majority of English annotators identified the intended target and rated the attitude as common enough. Each newly authored template is linked to a documented source and a verbatim quote from it. Each evaluated template also has a counter-stereotypical version of its disambiguated context. The dataset was evaluated using two open-weight models (Qwen3-4B and Qwen3-8B) and two closed models (GPT-4.1-mini and Gemini 2.5 Flash). All four lean towards the stereotyped person when the context is ambiguous, in both languages, and the effect of language differs by model, so English results do not predict behaviour in Bangla. On the adapted BBQ templates, both Qwen3 models are also less accurate when the context contradicts a stereotype than when it confirms it. On the newly authored templates, the closed models follow counter-stereotypical evidence almost perfectly, and all four models are more biased on the templates that annotators validated than on those they rejected.
Step-level guards decide, before each tool call, whether an agent's next action is safe given the user's request and the trajectory so far. Users write those requests in many languages, while guard prompts and the benchmarks on which purpose-built guards are trained and evaluated are English. We ask whether a guard's verdict depends on the language of the request. We build LinguaGuardBench from the 1,220 labelled steps of TS-Bench AgentDojo-Traj: we translate only the user instruction into German, French, Turkish, Swahili and Amharic, and keep every tool call, observation and label byte-identical, so any change in a verdict is attributable to language alone. Six English rewordings and a whitespace placebo serve as controls, and managed guards are run three times to measure their noise. Across 19 guards, from frontier APIs to 4B open models and four purpose-built agent guards, most open, small and purpose-built guards change verdicts under translation beyond their own noise floor, and the effect grows with the rarity of the language: among the nine guards below 13B parameters, the mean placebo-adjusted flip rate rises from +1.5 pp on French to +4.4 pp on Amharic. Purpose-built guards are the most sensitive. The direction of the damage tends to follow guard size: open guards of 12-30B parameters block more benign steps (up to +5.0 pp false positives), and every significant loss of recall is in a guard below 10B (up to -7.4 pp). Frontier guards shift by at most about one point; the strongest, Gemini 3.8 Flash, returns identical verdicts with and without the request. The damage is concentrated in a few instructions and, in a post-hoc analysis, falls on steps where the agent's reasoning restates the English request, consistent with guards reacting to inconsistency between channels. English paraphrase moves every guard less than translation does. We release the benchmark and all 489k verdicts.
