Math 2.0

Resources / preprint

Interpreting Reasoning of Large Language Models via Partial Information Decomposition W50NNC2MJS

URIm2://res/W50NNC2MJS
Kindpreprint
CreatorsBarproda Halder, Qiuyi Zhang, Sanghamitra Dutta
Date2026-09-30
Subjectsarxiv cs.IT, arxiv cs.AI, arxiv cs.CY, arxiv cs.LG, arxiv math.IT
Scopein

Expressions

e1: version, preprint

Languageen
Date2026-09-30
IdsarXiv 2610.00571

Items in this resource

The mentions of the inventory that name this resource. Each name links to the card of its inventory row.

ItemKindWhereSource
Monotonicity under local operations on $B$statementlemma 1arxiv
Unique informationdefinitiondefinition 1arxiv
Decomposition for repetitive reasoningstatementproposition 1arxiv
Trajectory Repetitive Reasoning Index (Trajectory-RRIdefinitiondefinition 3arxiv
Monotonicity under adversarial side informationstatementlemma 2arxiv
Step-RRI under Repetitive Reasoningstatementtheorem 2arxiv
Uniqueness under noisestatementtheorem 1arxiv
Step-wise Repetitive Reasoning Index (Step-RRIdefinitiondefinition 2arxiv

Catalogue records