Fig. 4: Experimental setups for previously unknown solutions exhibit comprehensible patterns.
From: Meta-designing quantum experiments with language models
In the top two rows, we show two previously unknown constructions for the spin-\(\frac{1}{2}\) states and the Majumdar–Ghosh states (described in more detail in ref. 20). For each of the two examples, the code produces the correct experimental setup for the three states used to prompt the model, as well as for higher particle numbers, indicating that the model was able to pick up on the pattern and write the correct code for the entire class of states. We highlight the ‘building blocks’ in green, which are repeated multiple times as the particle number grows (stemming from lines written in the for loop). The bottom row shows a code for the Dyck 1 state. The setups generated by this code produce the correct state up to the third iteration, but are missing terms for indices N > 2. This means that the model was able to solve the task it was trained to do (match the first three states), but failed at the meta-task of picking up on the pattern we intended it to match beyond the first three examples. It is also notable that in contrast to the other two examples, all setups produced for the Dyck 1 state also contained additional crystals that did not actually contribute to the resulting quantum state. We have omitted them by masking them by a grey rounded rectangle.
