Credit Attribution and Stable Compression

Authors: Roi Livni, Shay Moran, Kobbi Nissim, Chirag Pabbaraju

NeurIPS 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical The paper does not include experiments.
Researcher Affiliation Collaboration Tel Aviv University. rlivni@tauex.tau.ac.il. Technion and Google Research. smoran@technion.ac.il. Georgetown University and Google Research. kobbi.nissim@georgetown.edu. Stanford University. cpabbara@cs.stanford.edu.
Pseudocode Yes Definition 7 (Randomized response). Let RR : {0, 1}𝑛 [𝑛]π‘˜be the randomized response selection mechanism defined as follows. Given π‘₯ {0, 1}𝑛, RR flips each bit of π‘₯independently with probability 1 1+𝑒Ρ to obtain π‘₯. Let 𝑆= {𝑖 [𝑛] : π‘₯𝑖= 1} and 𝑆 = [𝑛]  𝑆. Further, let |𝑆| = 𝑑. If 𝑑 π‘˜, then RR outputs a uniformly random subset of π‘˜indices from 𝑆, ordered arbitrarily. Otherwise, it arbitrarily orders 𝑆, and outputs 𝑆 𝑇, where 𝑇is a uniformly random subset of π‘˜ 𝑑indices chosen from 𝑆 (and ordered arbitrarily), and denotes concatenation.
Open Source Code No The paper states in its NeurIPS checklist that it 'does not include experiments requiring code', implying no open-source code for the described methodology is provided.
Open Datasets No The paper is theoretical and does not report on experiments that would use a training dataset.
Dataset Splits No The paper is theoretical and does not report on experiments that would use a validation dataset.
Hardware Specification No The paper is theoretical and does not report on experiments, thus no hardware specifications are provided.
Software Dependencies No The paper is theoretical and does not report on experiments, thus no specific software dependencies with version numbers are listed.
Experiment Setup No The paper is theoretical and does not describe any experimental setup or hyperparameters for empirical evaluation.