Practical prompting, backed by research

An LLM gives different answers to the same question in Spanish and in Japanese. Ask about the capital of France in Spanish — one answer; in Japanese — another. The problem: the model doesn't realize that "Francia" and "フランス" are the same entity. It maps every concept into its internal English space, and that step is where it often slips. SUBSUB and SUBINJ let you get consistent facts in any language — you add the English translation of the key concept right into the query. The model skips the unreliable internal translation and works straight in its "native" English space — +5-44% cross-lingual consistency, +2-32% factual accuracy.

If an LLM edits your text and the editor is still unhappy, the model isn't the problem. Every editor has their own unwritten playbook. It lives only in their edits: 'too formal', 'take the discount out of the headline', 'add numbers'. The Feedback-to-Rubrics method lets you extract these rules from accumulated comments and turn them into an explicit list of criteria for evaluating new texts. The trick: the LLM doesn't just read the edits, it tests hypotheses. It predicts what comment this criterion should have produced. If that diverges from the real one, it refines. After 2-3 rounds, a chaotic stack of edits becomes a working list of requirements that works without its author.

Giving an LLM a perfect worked example delivers only 60-70% of the benefit. The other 30-40% is lost because the model doesn't see the structural pattern — it latches onto surface differences ("investing" vs "marketing") and misses that the logic is the same. SA-ICL tackles knowledge transfer between tasks that look different but share a structure. The method builds an explicit schema of the task (category → details → focus → goal), finds a similar one from past experience and activates it — merging the retrieved pattern with the current problem. The model stops guessing at the connection and starts working from a ready-made template.