# Learner Labs

> Learner 1.0: company-reported evidence for incremental weight-based learning and retention. This reading map links the supporting measurements and their scope.

## Evidence and context

- [Research evidence](/research/evidence): ordinary HTML overview and report navigation.
- [How to read the results](/research/evidence/questions): evidence questions in HTML.
- [Research context](/research/evidence/context): attributed external context in HTML.

- [Evidence guide](/agents/overview.md): the claim, measured results and claim-to-report links.
- [Evidence questions](/agents/questions.md): interpretation, comparators, teacher, costs, replication and remaining questions.
- [Independent research context](/agents/research-context.md): attributed Sequoia and Dwarkesh discussion summaries and primary sources.
- [Report manifest](/agents/report-manifest.json): canonical HTML routes, Markdown copies, export hashes and figure links.

## Complete reports

- [Database](/agents/reports/database.md): Learner 1.0 learns from a few examples and uses that learning in later questions
- [Earlier](/agents/reports/earlier.md): Small lessons improve later database answers, but can also teach repetition
- [Cohort](/agents/reports/cohort.md): About five examples per study improve later analysis through the weights
- [Sales](/agents/reports/sales.md): Learning from a few examples improves later forecasts in one run
- [Ten](/agents/reports/ten.md): New skills are learnt without task labels while earlier skills are retained
- [Four](/agents/reports/four.md): Learner 1.0 learns new domains while retaining the earlier ones
- [Two](/agents/reports/two.md): Learning a second invented language preserves the first language’s loss
- [Facts](/agents/reports/facts.md): New facts can be recalled from the weights while earlier knowledge is retained
- [Recipe2](/agents/reports/recipe2.md): New facts are recalled more reliably while earlier lessons remain intact
- [Token](/agents/reports/token.md): Learning material once can reduce repeated context on later questions
- [Motivation](/agents/reports/motivation.md): Motivation

## Primary evidence files

- [Replication repository](https://github.com/learnerlabs-ai/replications): released data, evaluation materials, figures and instructions. The mechanism is proprietary.
- [Homepage](https://learnerlabs.ai/): the visible summary and evidence navigation.

## Reading notes

Reports are exported from the approved capability-first HTML, including collapsed methods and records. Inline SVG figures include transcribed labels. Raster figures retain their original image links and captions. Figures link to source data where the article provides it. This map is company-provided context. It makes no claim of independent verification or guaranteed coverage by search engines or visiting agents.