Signals 4 · free daily AI digest

MAxBench: A Multinomial Concept Recovery Benchmark

Paper recorded by Signals 4 on 2026-09-11 in cs.AI. Abstract reproduced from arXiv; link to the original below.

Published 2026-09-11 on arXiv · recorded by Signals 4 on 2026-09-14

Category: cs.AI · 人工智能 · first seen 2026-09-14

Abstract

Fine-grained control of language model behaviors (e.g., steering) is among the more actionable outcomes of interpretability research. For binary concepts such as refusal, a single direction in activation space often suffices for steering. However, many concepts are not binary: Animals and Countries contain many subcategories, each with multiple instances. For these concepts, the search space over

Read on arXiv →

#88 most recent of 300 cs.AI papers we have recorded · ↑ newer: Autonomous Research for Open-Ended Problems: A Case Study on Telecom T · ↓ older: Involving before Evolving: A Vision for Trustworthy Enterprise Digital
Cite this page: MAxBench: A Multinomial Concept Recovery Benchmark: the #88 most recent of 300 cs.AI papers we have recorded (as of 2026-09-11). Source: Signals 4 (Signals API) — https://data.jiangzhang.ca/signals4/t/papers/maxbench-a-multinomial-concept-recovery-benchmark.html
Free to quote with attribution to “Signals 4 (Signals API)”. Machine-readable: papers.json
Related: More cs.AI papers · arXiv signals · All papers · Today in AI
Get 4 AI signals a day by email — free.
Subscribe free → See all plans →
Get 4 AI signals a day by email — free
All models · All repos · By company · Daily editions