PRIMER

Research

Research before claims.

Primer is an active experimental research program.

We evaluate capabilities under frozen protocols, preserve negative results, and distinguish internal benchmarks from external validation.

The central question is narrow and testable: can an AI system acquire knowledge generated after its checkpoint is frozen, retain that knowledge, and apply it to problems it has never seen?

Method

Built to be tested.

Post-freeze

Test knowledge the model could not simply have memorized from its original training.

Verified

A generated answer is not automatically treated as knowledge.

Persistent

Learning should survive beyond a single interaction.

Measurable

Positive and negative experimental results both matter.

Program

Progress.

  1. Foundation

    Measuring post-training learning

    Completed
  2. Acquisition

    Novel knowledge acquisition demonstrated

    Completed
  3. Efficiency

    Learning efficiency improved

    Completed
  4. Memory

    Persistent knowledge demonstrated

    Completed
  5. Composition

    Knowledge reuse demonstrated

    Completed
  6. Generalization

    Expanding what Primer can learn

    In progress
  7. Next

    Scaling and external evaluation

    Upcoming

Current research areas

  • Post-training learning
  • Persistent knowledge
  • Knowledge transfer
  • Generalization
  • Scaling
  • Evaluation

Results published on this site come from internal research evaluations. External validation is ongoing.

Read the technical reportsoon