INDEX

Explanations

Refers to items labeled "A" (or similar)

np_max-act · gemini-2.0-flash

This neuron detects placeholder entity tokens of the form “NAME_<digit>” (e.g. NAME_1, NAME_2).

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

reminder

-0.07

.frames

-0.07

 döndü

-0.06

(bodyParser

-0.06

 GameState

-0.06

 ventured

-0.06

 paci

-0.06

 fandom

-0.06

gö

-0.06

sve

-0.06

POSITIVE LOGITS

verting

0.08

 advancing

0.06

 lớn

0.06

uting

0.06

�

0.06

ˆ

0.06

xxx

0.06

 instantiated

0.06

 oriented

0.06

 subsequently

0.06

Activations Density 0.025%

Refers to items labeled "A" (or similar)

This neuron detects placeholder entity tokens of the form “NAME_<digit>” (e.g. NAME_1, NAME_2).

No Comments

No Known Activations