Loading
For teams training and fine-tuning frontier models
Static-corpus models hallucinate dated facts and saturate on benchmarks that leaked into their pretraining.
NOSIBLE WORLD is a dated, multilingual, replayable archive that fixes both.
A frontier model trained on text that ends in October still answers questions asked in March. It either hallucinates a dated fact, or recites a benchmark answer that leaked into its training corpus. Dated retrieval fixes both.
Which 1968 novel introduced the term “replicant”?
Do Androids Dream of Electric Sheep?
The question and the answer both appear verbatim in the pretraining set, so the model is reciting from memory rather than reasoning from evidence.
As of 2026-04-15, what fraction of Humanity’s Last Exam can the top open model solve?
Resolves to events published before 2026-04-15. Cites each source by publication minute.
The answer resolves to a record published after the eval snapshot. Replay the prompt at any past as_of date and the answer changes accordingly.
Frontier LLMs lag humans on temporal reasoning, and static benchmarks now leak into pretraining. The fix is consistent across these papers: pin the training cutoff and score on events that resolve after it.
If your training corpus ends in October, your model lives in October. The world does not.
One hundred million events mined from the open web, each one carrying a verified first-publication timestamp, persistent actor identifiers, full source evidence, and labels from seven independent ontologies.
EU AI Act enters into force, imposing dataset and copyright disclosure on foundation models.
OpenAI announces o3 with frontier gains on ARC-AGI, resetting the reasoning benchmark frontier.
DeepSeek-R1 open-weights release reprices US AI majors on a single Monday.
Bartz v. Anthropic settles for $1.5B over roughly 500K pirated training-corpus titles.
NYT v. OpenAI: court orders production of 20M ChatGPT logs to plaintiffs.
Humanity's Last Exam climbs from 10% to 46% in twelve months; static evals saturated.
Static corpora start aging the day they ship, and the models trained on them inherit the date. NOSIBLE WORLD is the fix: an open-web archive with every event dated to the publication minute, replayable to any past as_of.
Pretrain on the open web, dated to the minute.
Each one wires into your existing training and evaluation stack.
A chronologically ordered token stream where every document carries a verified first-publication timestamp. Replay it byte-for-byte at any past as_of date your eval requires.
Pin a training cutoff, score against events that resolve after it. Forward-window questions grow with the ledger. Compatible with the ForecastBench and AntiLeak-Bench protocols.
Instruction-response pairs where every cited fact carries its publication time, source, and language. The model is trained to refuse when the evidence post-dates its corpus.
NOSIBLE WORLD: a dated, replayable corpus for training and evaluating point-in-time LLMs.
