Roberto I. Ono Filho

Independent researcher, São Paulo. Machine novelty: what stimulating a language model buys, measured.

Where the stream goes (paper 1). One opening sentence, “Her brother collected sounds the way other children collected stones.” (★), continued by a base model in two ways. Each dot is a window of 64 generated tokens, embedded in sentence space and projected on the same plane in both panels, each panel drawn to its own scale; the trail follows the order of generation. Left alone, the model makes one jump and freezes into repetition. With a new subject injected every 150 tokens (accent), it keeps crossing the space. Redrawn from the paper’s data, one seed; the effect on judged surprise and connection is measured over ten premises in the paper.

In one paragraph

Is the answer to an open problem already inside a language model, waiting for the right stimulus? I have been trying to find out at office scale, one laptop and rented GPUs by the hour, under one discipline: pre-registered batteries read once, blind judges that score only generated text, exact verifiers with classical baselines, sham controls, every candidate logged. So far every inference-time stimulus bought variation or integration and none bought progression; moving the weights moved the mass toward the known good and paid with the tails; and in verified search every proposer stalls at the same plateau, a record away. Negative results are published with the same care as positive ones.

Papers

  1. 04

    Pulsed, Not Pressed: What State-Space Injection Buys in Open-Ended Generation, and What It Costs

    Zenodo, doi:10.5281/zenodo.22280784 · September 2026

    Intervening in the residual stream during generation. Pulsing a concept direction for a few tokens every few hundred raises judged surprise where pressing it continuously collapses coherence; the direction's content carries the effect (shams fail); against the written subject, the pulse matches text interruption at its peak, trails it everywhere else, and costs coherence.

  2. 03

    Operator Packages, Proposer Strength, and Construction-Family Plateaus in Office-Scale Verified Search

    Zenodo, doi:10.5281/zenodo.22280391 · September 2026

    A FunSearch-style loop instrumented at office scale, on nine construction problems from a public repository. Three operator packages compose additively and close more of the seed-to-record gap; a frontier proposer reaches in tens of samples what the local model does not in hundreds; and every proposer stalls at the same plateau, a record away; a later analysis reads the flagship wall as the grid's resolution limit rather than the family. Named in words, the record's family is adopted and loses; handed as code, it is optimized.

  3. 02

    Moving the Mean Toward the Known Good, Not Beyond It: What Inference-Time Interventions and Weight Consolidation Buy in Open-Ended Generation

    arXiv:2608.28886 · August 2026

    Every inference-time intervention bought variation or integration and none bought progression; a verifier placed inside the flow was imitated rather than obeyed. Consolidating the model on its own verified findings moves what it writes toward the known good, replicated in three lineages, and never beyond it: the best candidate stays the classical heuristic, and the tails where better-than-classic candidates live get thinner.

  4. 01

    Interrupting the Loop: Periodic Subject Changes Raise Judged Surprise and Connection in Base Language Models

    arXiv:2608.19893 · August 2026

    A base model writing with no task drifts into repetition. Of the operations usually credited for novelty, one survives measurement: a new subject injected every few hundred tokens into a habituated stream raises judged surprise and connection, replicated across models and on held-out premises. The anti-probable sampler and the improbable prompt do not.

Method

Code and data

github.com/RobertoOno/interrupting-the-loop: the harness (samplers, the reverie loop, state-space operators, the verified-search loop and its domains), the analysis scripts, the papers’ LaTeX, and the lab notebook. Run data as releases: run-data-v1 (papers 2 and 3) and run-data-v2 (paper 4).

Contact

Email
roberto@robertoono.com
ORCID
0009-0006-8650-629X
GitHub
RobertoOno