INDEX

Explanations

inscription

np_max-act · gemini-2.0-flash

This neuron detects references to carved or engraved monument text (e.g. “inscribed,” “inscription,” “etched”).

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

 Love

-0.07

-rock

-0.07

 barcelona

-0.07

OOK

-0.06

Time

-0.06

 deck

-0.06

 tower

-0.06

 ball

-0.06

 rock

-0.06

-mar

-0.06

POSITIVE LOGITS

 inscription

0.08

 guiActive

0.06

сти

0.06

=settings

0.06

 дан

0.06

 engraved

0.06

 hissed

0.06

.getOwnProperty

0.06

 그리고

0.06

 preserved

0.06

Activations Density 0.004%

inscription

This neuron detects references to carved or engraved monument text (e.g. “inscribed,” “inscription,” “etched”).

No Comments

No Known Activations