INDEX

Explanations

the

np_max-act · gemini-2.0-flash

This neuron activates on normative language stating broad rules or defaults (e.g., phrases like “general rule” or “default deny”).

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

 INCLUDING

-0.06

WK

-0.06

 convinc

-0.06

เคล

-0.06

Na

-0.06

Ger

-0.06

olit

-0.06

 originating

-0.06

VGA

-0.06

 문서

-0.06

POSITIVE LOGITS

ΟΜ

0.07

↵↵↵↵↵↵

0.07

 Niagara

0.07

 mains

0.07

中华

0.07

 Arizona

0.07

 conditional

0.07

 Sunshine

0.06

aret

0.06

��

0.06

Activations Density 0.014%

the

This neuron activates on normative language stating broad rules or defaults (e.g., phrases like “general rule” or “default deny”).

No Comments

No Known Activations

the

This neuron activates on normative language stating broad rules or defaults (e.g., phrases like “general rule” or “default deny”).

No Comments

No Known Activations