Planning or Improvisation? Stress-Testing the Poetry Planning Site on Open Models and Open Cross-Layer Transcoders

Abstract:Lindsey et al. (2025) report that Claude 3.5 Haiku plans rhymes: features for candidate rhyme words are active on the newline before a line is written, and a suppress-and-inject intervention redirects the line only when applied there (their Figure 13). We test how far this generalizes on seven cells crossing four open models (0.6B to 2.6B parameters) with six open cross-layer transcoders (CLTs), on one consumer GPU, decomposing the claim into position specificity (C1), newline site identity (C2), and a newline-resident plan (C3). This is a stress test rather than a faithful reproduction: attribution graphs are unavailable for these CLTs, so features are found bottom-up from decoder vectors. C1 generalizes, in every cell and in all 247 of 444 prompt-by-inject pairs with a detectable effect, but the effective position is the final prompt token, adjacent to emission, and only two cells reach behaviorally meaningful probabilities. C2 and C3 are not recovered by any probe: a census of every active feature finds no rhyme-anticipating enrichment at the newline, and steering the newline while the model composes the whole line, over 36 runs and 8,640 sampled lines, shows why. That intervention is strong but one token long, making the injected word the first word of the composed line in 703 of 720 samples and leaving the rhyme six words later untouched. A final test drops the transcoder entirely: patching the newline's whole residual, at every layer, from a minimal-pair poem whose third line ends on a different rhyme moves the rhyme in 11 of 1,260 composed lines against 4 at baseline, with a design resolving 1.4%. We read this as a boundary condition rather than a refutation: at this scale and with these transcoders, the causal site is emission-adjacent. We reproduce Figure 13's shape, not its mechanism. Code and data are public (code: this http URL).

17 pages, 3 figures, 8 tables. Code, data and analysis scripts: this https URL . An earlier version was submitted to the BlackboxNLP 2026 special track on reproducibility and reliability in interpretability analyses; this version adds a rerun composition-horizon experiment (36 runs, 8,640 sampled lines) and a transcoder-free activation-patching test

Computation and Language (cs.CL)

arXiv:2609.18440 [cs.CL]

(or arXiv:2609.18440v1 [cs.CL] for this version)

https://doi.org/10.48550/arXiv.2609.18440 arXiv-issued DOI via DataCite (pending registration)