Research
Independent research at the edge of translation, world-model agents, and post-transformer attention — each built from first principles on local hardware, with results reported straight, including the negative ones.
Three lines of work, three different bets. What ties them together is method: falsifiable questions, instruments built rather than assumed, honest numbers, and a clear line between what is finished enough to show and what is held back until formal publication.
Can 'untranslatability' be measured? The three-way honorific register of Bengali and Hindi is grammatically forced but vanishes in English, so any translate-through-English pipeline must guess it back. A rule-based register classifier (precision ≥ 0.90 per class, both languages) is the instrument; the measurement quantifies the obstruction per sentence and tests, under pre-registration, whether it predicts downstream failure. Method and dataset held back pending arXiv.
Three independent agents that induce a game's hidden rules online, probe under uncertainty, and plan in an internal simulator. Rigorous offline evaluation (replay verifiers, held-out log pairs, baselines); modest live scores reported straight — best run 4/183 levels, aggregate ≈ 0.257% — with the exact broken invariants diagnosed rather than hidden.
A sub-quadratic O(n log n) structured replacement for self-attention, benchmarked forward-only on real Gemma-4-E2B and Llama-3.2-1B weights. Faster than dense attention past a ~1330-token crossover and monotonically further ahead out to 14k+ tokens. Reported as a speed/scaling result on an untrained gate — the efficiency characteristic, not modeling quality; mechanism held back pending write-up.
01
02
Negative results are results: a zero live score with an exact root-cause diagnosis, or a fast mechanism whose untrained accuracy collapses, are reported as they are. Where ground-truth labels came from a model rather than a human, that is disclosed, because it changes what the numbers mean.
Priority note. The unpublished method and frozen dataset behind Register Obstruction are deliberately not in the public repositories, to preserve priority ahead of peer review. Happy to walk through the full work in conversation.
Get in touch about the research