<b>Meaning in the Reader: How Language Models Assemble Scientific Assessments:</b><b>Factor-Specific Signals, Local Semantic Binding, and Interface-Dependent Readout</b>

Third, exact-state interventions<br>identify a clause-local causal assessment signal. Consequence-associated perturbations<br>transfer across scenes and paraphrases, but their semantic field is not self-contained in<br>the donor state. A perturbation derived from another factor can produce nearly the same<br>consequence effect when inserted into the recipient consequence slot, whereas the same<br>consequence perturbation is much weaker when inserted into a different slot.

The donor<br>directions themselves are only moderately aligned. These results support a reader-assigned<br>assessment computation: upstream states preserve factor-sensitive signals and partially<br>portable signed/value components; the recipient slot and local reader assign much of the<br>field semantics; and the report interface determines how these components are mixed and<br>amplified into settlement, consequence, verification, relevance, and other outputs.

The earlier<br>low-dimensional assessment geometry is therefore better understood as a reader-specific<br>pooled summary than as a universal fixed semantic coordinate system. Scientific judgement<br>in language models is not simply stored and then reported; it is assembled at readout</p>