A retrieval-augmented encoder-decoder that injects 'in-context vectors' into latent states is presented, with claims of competing with much larger RAG models on three QA benchmarks.
Entropy of the quantum work distribution
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
abstract
The statistics of work done on a quantum system can be quantified by the two-point measurement scheme. We show how the Shannon entropy of the work distribution admits a general upper bound depending on the initial diagonal entropy, and a purely quantum term associated to the relative entropy of coherence. We demonstrate that this approach captures strong signatures of the underlying physics in a diverse range of settings. In particular, we carry out a detailed study of the Aubry-Andr\'e-Harper model and show that the entropy of the work distribution conveys very clearly the physics of the localization transition, which is not apparent from the statistical moments.
fields
cs.CL 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Efficient Knowledge Feeding to Language Models: A Novel Integrated Encoder-Decoder Architecture
A retrieval-augmented encoder-decoder that injects 'in-context vectors' into latent states is presented, with claims of competing with much larger RAG models on three QA benchmarks.