INDEX

Explanations

Mentions of time and numbers

np_max-act · gemini-2.0-flash

This neuron selectively activates on standalone header or title tokens—e.g. section headings (like “History,” “Cultural impact”) and isolated numbers or labels that mark the start of a new section.

oai_token-act-pair · o4-mini Triggered by @xinyanhu8

New Auto-Interp

Configuration

andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_11/trainer_1

Dataset (Dashboard)

Various

Features

131,072

Data Type

float32

Hook Name

blocks.11.hook_resid_post

Architecture

standard

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ge

-0.06

dehyde

-0.06

regular

-0.06

ав

-0.06

Esk

-0.06

angstrom

-0.06

 principal

-0.06

>t

-0.06

 어떻게

-0.06

Microsoft

-0.06

POSITIVE LOGITS

.ToDecimal

0.07

 Nack

0.06

_classification

0.06

NullException

0.06

 passion

0.06

\Schema

0.06

 passionate

0.06

_grupo

0.06

 characterized

0.06

싱

0.06

Activations Density 0.043%

Mentions of time and numbers

This neuron selectively activates on standalone header or title tokens—e.g. section headings (like “History,” “Cultural impact”) and isolated numbers or labels that mark the start of a new section.

No Comments

No Known Activations