INDEX

Explanations

incidents involving actions that can cause physical harm or distress

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 15-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_15/width_16k/average_l0_78

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.15.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

olg

-0.51

 Background

-0.49

paña

-0.49

feer

-0.47

 nahilalakip

-0.47

 Vort

-0.46

orie

-0.46

 background

-0.45

Ag

-0.45

 donate

-0.45

POSITIVE LOGITS

complexContent

0.78

addContainerGap

0.66

 arrival

0.66

ViewFeatures

0.65

 arrives

0.65

 arrive

0.62

UnknownFieldSet

0.62

 contextLoads

0.61

 arrivals

0.60

 useStyles

0.60

Activations Density 0.223%

incidents involving actions that can cause physical harm or distress

No Comments

No Known Activations