Veritate

Research log for Carpathian's byte-level, efficient AI inference engine

6 Articles

Long-Context Memory For Constant-State Models

The conversation-needle benchmark, the transfer-gap finding, bitwise-exact streaming state carry, and the open programme for training past-window recall.

Jul 9, 2026

20 min read

How A 121M Byte Model Learned To Talk

The three-phase recipe behind chat80m, the first conversing byte-level model in the project, its knowledge wall, and the identity-dosing discovery.

Jul 9, 2026

20 min read

The Composed Efficiency Stack

Byte-level training efficiency on one consumer machine: Muon, boundary patching, constant-state recurrence, and the hybrid trunk that beat both of its parents.

Jul 9, 2026

19 min read

Amortizing The Optimizer Step

How an affine, fixed-cost-per-update model of training time turned a budgeted four-to-five-day run into a measured two-day run.

Jul 9, 2026

22 min read

Serving A Research Architecture On CPU

The v13 engine: Constant-size decode state, byte-for-byte parity to the reference and about one millisecond per byte on a CPU core.

Jul 9, 2026

18 min read

Negative LLM Training Results

Falsified approaches at 10M+ params: fast-weight mem, test-time depth, byte-level selective loss, delta-rule instability, mem registers, and the MoE tax.

Jul 9, 2026

30 min read