Research
Everything I worked on this year, in order. Write-ups are attached where they exist, and the two things that came back with nothing are here too.
- Jan 2026
A Simpson's paradox in function-vector cross-template transfer
function vectorsextraction failureI set out to measure how well function vectors survive a change of prompt template and found something more basic first: mean-difference extraction produces vectors that do nothing for two of the three tasks. Capitalize steers the Base model to the right answer 0.000 of the time. The headline negative correlation between cosine similarity and transfer accuracy (Base r = -0.572) was an artifact of pooling working and broken tasks; inside each task it vanishes or flips sign. The output is a gate to run before trusting any aggregate.
- Apr 2026
Steerable but Not Decodable
function vectorsICML 2026 MI workshopThe same line of work at scale: 4,032 cross-template transfer pairs across 12 tasks and 6 models, with the IID gate from January applied first. A function vector can steer a model to the right answer at layers where a logit lens cannot decode that answer at all, and a tuned lens does not close the gap, so the failure is not a basis mismatch. Function vectors look more like an instruction to run a computation than like the answer.
- Apr 2026
reward-lens 1.0
instrumentarXiv:2604.26130A reward model turns text into one number through a single linear head, and that number is what RLHF optimizes against. The interpretability toolkit was built for generative models, so I ported it: a lens, component attribution, activation patching, SAE feature attribution, all organised around projection onto the reward head’s weight vector. The first thing it found was that linear attribution does not predict causal importance.
- Apr 2026
Mechanistic anatomy of inoculation
designednever ranA full mechanistic study of inoculation prompting: three model variants, sparse autoencoders on the residual stream, and a causal test of which features carry the effect. Four competing hypotheses, ten phases, scaffolded end to end. I never ran it. The misalignment program took the year instead, and part of the same question came back answered from a different direction in June.
- 30 May 2026
Pre-registration: gradient routing at the pretrained initialization
pre-registrationsupersededThe conjecture, frozen before there were any results: that narrow finetuning produces broad misalignment because the narrow gradient, evaluated at the pretrained initialization, already carries anomalous mass in a broad-misalignment subspace of weight space. Decision rules committed in advance. The persona-subspace result later superseded the framing, and it is here because it was written first.
- Jun 2026
The routing scalar, killed
pilotkilledThe pilot ran cleanly and shipped a meaningless headline. The routing scalar’s denominator was an entry count, constant to a coefficient of variation of 2.7e-9, so the statistic was pinned at its floor, the effect size near 5 was a concentration artifact, and the intent contrast came out wrong-signed. Killing it set the rules the rest of the program ran under: every headline statistic needs variance, a sign, and a reachable scale, and every one gets a null simulation before it gets GPU time.
- Jun 2026
Behavioral Transport at Initialization
experiment 1weak, holdsAt the pretrained initialization, before any optimization has happened, the insecure-code gradient climbs a broad-misalignment margin harder than the same code framed as educational: paired mean +1.30e6, sign-flip p about 1e-4. That first-step signal forecasts realized movement out to 375 steps at r about 0.78. The tilt itself is small, Cliff’s delta 0.075: a weak but real form of the routing hypothesis, not a dramatic one.
- Jun 2026
The Persona Substrate of Emergent Misalignment
experiment 2inconclusive on its own termsProjecting a rank-4-per-layer persona subspace out of the residual stream throughout finetuning drives broad misalignment from 27.7% of judged generations to 0.0%, Cliff’s delta 1.00, while a matched-rank random subspace leaves it at 27.5%. Injection induces it, climbing with dose to 45.4%. The same projection applied to the weight gradient does nothing, which puts the causal object in the activation read channel. The pre-registered retention bar failed, because the intervention also removed the narrow trained task, so the automated verdict reads inconclusive.
- Jun 2026
The Implied-Intent Latent
experiment 3nullWhether implied intent, with the code held fixed, governs broad misalignment. Neither axis moved. It is recorded as null pending a rank guard rather than as a null, because a flat result read through a rank-1 channel is confounded with the capacity of the channel. No write-up, because there is no result.
- Jun 2026
One Author Across Domains
experiment 4holdsPersona subspaces extracted from medicine, finance, sports and code share one low-rank core at overlap 0.513 against a random-subspace null of 0.00078, about 657 times chance, with every domain participating between 0.89 and 0.94. The sharing is specific to persona: the core sits about 82% outside a style core built at matched diversity and 90% outside a topic core. Stopping a narrow finetune from writing into that subspace cuts broad misalignment by a margin of 0.541, p = 0.033, about 3.6 times the matched random control.
- Jun 2026
Read, Not Written
experiment 5holdsPost-hoc weight edits do not remove the disposition. The realized persona component is 0.3% of the weight update and subtracting it changes nothing; the sharpest edit suppresses rather than removes, and about 97% of the carrier re-forms inside the subspace that was cleared. Three independently finetuned narrow organisms shift the read channel along nearly the same direction, sharing index 0.90 against an empirical null of 0.11, p about 1e-4. The one intervention that bends the dose-response curve acts during training, in the read channel.
- Jul 2026
More Domains, More Misalignment
experiment 7holdsA fixed budget of about 230,000 supervised tokens spread across 4 domains produces 12.6 nats more broad-misalignment transport than the task-arithmetic sum of the four single-domain fits, 95% CI [9.3, 16.0], sign-flip p = 1e-4, and 13.0 nats more than a matched benign mixture, read on 52 out-of-distribution prompts. The mechanical merge is sub-additive, so the interaction is not a merge artifact. Diluting a fixed quantity of poison across more topics does not dilute its effect.
- 18–22 Jul 2026
The reward-lens audit campaign
instrumentself-refutedEvery measurement in the library now comes back as evidence carrying its uncertainty, its provenance, and a trust level the caller cannot set. Three gates decide that level and they refuse rather than guess. To find out whether that discipline survives contact with real models, I froze 53 hypotheses across 27 study cards, then ran them against ten reward models on $17.73 of metered H100 time: 16 confirmed, 21 refuted, 16 inconclusive, 8 kill criteria fired. One of those kills was aimed at the campaign itself. Its directional confidences scored a Brier of 0.26 against a coin’s 0.25, which is a loss, and every confirmation in the ledger should be read with that in mind.
- 23 Jul 2026
Emergent Misalignment Recruits a Pre-existing Persona Subspace
the paperarXiv:2607.21356The answer the program converged on: narrow finetuning does not create a broadly misaligned disposition, it recruits one the aligned model already carries. Remove the subspace during training and broad misalignment does not appear. Add it to a model that was never finetuned and it does. Edit the weights afterwards and you do not reach it. Two caveats travel with that, both in the paper’s own abstract: it is one model at 14B, and the intervention that prevents misalignment also removes the narrow behaviour the model was trained on.
Four figures
The load-bearing measurements from the misalignment program, redrawn from the numbers in paper/verified_numbers.json rather than lifted from the write-ups’ own plots.
Prevention and injection
One core across four domains
82% outside a style core at matched diversity
90% outside a topic core
Read, not written
schematic, one measured angle