A preprint is a signal, not a finished product or an independently confirmed result. We therefore track research papers separately and show the actual state of supporting evidence for each one.
1381
published research events
Latest work
A significant claim from a single source is published only after further confirmation.
The study introduces SteerScope with 15 metrics and compares 23 methods. According to the authors, none of the evaluated activation steering methods outperformed Prompt Steering in effectiveness without greater overall side effects. For inputs outside the original distribution, side effects often intensified.
HARP continually reallocates unused budget during LLM evaluation. The authors prove that the method stays within a fixed resource limit and establish statistical guarantees for the estimates. Experiments cover task completion by agents, jailbreaking, toxic content and hallucinations when using RAG.
The study introduces HARP, which estimates the preferences of individual agents using explicit probabilities updated by Bayes' rule. The HARP+ variant supports steps that refine these estimates and, according to the authors, achieved the best results among methods without knowledge of the true preferences.
The WASD method accounts for semantic similarity between tokens during LLM distillation using their embeddings and Sinkhorn divergence. According to the authors, it improves performance on instruction following, mathematical reasoning and code generation. The implementation is publicly available.
A study involving more than 1 000 runs on ARC-TGI found that small language models can achieve high success rates within the training distribution, but their performance drops sharply outside it. The decline also occurs when the grid size changes while the rule remains the same.
The authors investigated a combination of knowledge distillation and quantization for biomedical translation from French into English. According to their experiments, model size decreased by 69 %, inference speed increased by 98.21 % and CO2 emissions fell by 98.46 % without any loss in translation quality.
The EMODE model separates the meaning of speech from emotional and prosodic signals and uses three-stage training. According to the authors, it improves both meaning preservation and sensitivity to emotions. The available abstract does not report numerical results.
A study using JNU data, with 1000 rpm excluded from training, measured classifier ensemble accuracy of 40.64 %, below the majority-class baseline. The standalone Logistic Regression model achieved 57.91 %, and its confidence estimate outperformed the proposed uncertainty decomposition.
The authors propose the State over Tokens framework: in their view, reasoning tokens preserve computational state between generation steps, but their text may not faithfully explain the model's reasoning. The available abstract presents the concept and directions for further research.
The authors of a commentary in Nature Machine Intelligence recommend testing whether using AI to present news emotionally helps spread accurate information or deepens divisions in society. The available excerpt does not include results from such testing.
The team at Lincoln Laboratory Supercomputing Center compares more than 120 AI accelerators in the latest LAICS study; the first study covered 57. The overview draws on publicly available data on peak performance and power draw, which some companies do not disclose.
Nitin Verma describes a study in which six models from OpenAI evaluated 137 anonymized manuscripts from the ACL 2017 conference. Each model produced three reviews of each manuscript following ACL guidelines. The available excerpt does not include the results.
The study authors trained eight algorithms on monthly data for four of the Great Lakes from 1982–2022. According to their results, the four best-performing algorithms had an error of about 12 cm compared with 14–21 cm for the simplest models. The SHAP method explained the contributions of the input variables.
Google presents five case studies of the use of the PDFM model in public health. According to the authors, its geographic representations achieved results equal to or better than conventional inputs without task-specific adjustments. The available excerpt does not contain detailed results.
A team led by PhAI Labs introduced JEPA-Anything, which combines several component predictions into a single prediction. According to the authors, in a test using a simplified Pong environment, it reduced error by 35 % compared with standard JEPA; for unseen combinations of interventions, it reduced error by 13 %.
A study of ten Code LLM models with 3–13 billion parameters found vulnerabilities of at least medium severity in 77.4–92.3 % of outputs generated with planted instructions in comments, compared with 1.7–5.1 % without them. The most effective follow-up check missed approximately a third of vulnerable outputs.
The study introduces ASCENT, a method that continuously updates agent weights using LoRA and verified task trajectories. The authors report higher success rates and efficiency in ALFWorld, WebShop and AppWorld without external example solutions.
The study introduces the JEVal benchmark with 11 257 examples and compares 25 model configurations. According to the authors, decision-making models can handle tasks with available evidence but are overconfident; in multi-step tasks, errors accumulate and reduce success rates.
AI Radar monitors Czech and international sources every day, looking for changes that truly deserve attention.
MonitorsOfficial AI company blogs, specialist media, and research sources.
Selects and combinesFilters out information noise and combines articles about the same change into a single event.
Summarizes and explainsExplains significant events in English: what happened, why it matters and where the information comes from.
The result is a quick overview of what has actually changed in the AI world, rather than another stream of articles.
Use the CS/EN switch to read the same Radar in Czech or English. English content is published after its translation has been checked, so new and older items may appear later.
Everything you need to navigate the AI world
Today’s briefingAn overview of news from the past 24 hours and a main story when available. Beside them is Live · AI Flash, followed by research and links to other Radar sections. On mobile, the blocks appear one below another.
AI FlashAn ongoing feed of brief updates with an evidence status. Links lead to a Radar detail page when one is ready, otherwise to the original source. You can also find reset and outage histories here.
Practical applicationsWhat new tools and features can do, what you can try and what their actual impact could be.
Model selectionModel comparison by type of work, capabilities, price and speed.
Research and archiveA separate research overview, topic search and older events by date.
One event, everything that matters
Each row represents one event — not one article. At a glance, you can see its significance, credibility and main point.
Illustrative example, not a current news item.
Importance: ▮▮▮ majorOpenAIModels✓ 6
Agent mode is available to all paying users
Until now, the mode was available only on the highest plan; it is now available on all paid tiers without a waitlist.
▮▮▮ major · ▮▮ important · ▮ we're tracking = how significant the change is✓ 6 = six independent publishers, not the number of articles or feeds✓ official = a clear release, law or incident is substantiated by the relevant authority1 source = no independent confirmation yetbold = who is behind the changegray text = a brief summary of what happened
The detail page contains a fuller summary, its significance and original sources. Practical impact appears in the detail and the For individuals and For businesses views. An AI Flash item reaches the main selection only after it has been expanded and meets the publication rules.
The same news, two practical uses
We first summarize each event in the same way for everyone. Based on those same facts, we then explain what the change means for your own use and what it could mean for how a company operates.
For individualsWhat you can use or try, how the change can help you at work and what to watch out for.
For businessesWhat impact the change could have on processes, costs, risks and other business decisions.
Today’s briefing is the same for everyone. Pages
For individuals and For businesses
can be found in the main navigation — they select only events relevant to the given use case.