INDEX

Explanations

questions expressing doubt or inquiry about circumstances and conditions

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

Followed by auxiliary verbs ("do," "did," "does")

np_acts-logits-general · gemini-2.0-flash

questions and "nor" clauses

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

how

-0.84

 what

-0.82

what

-0.73

 What

-0.73

How

-0.68

 whether

-0.67

how

-0.65

 they

-0.65

he

-0.64

whether

-0.60

POSITIVE LOGITS

do

1.29

did

1.15

 does

1.04

 تضيفلها

0.90

 saites

0.75

DO

0.71

do

0.68

DID

0.68

">—

0.66

WebVitals

0.66

Activations Density 0.164%

questions expressing doubt or inquiry about circumstances and conditions

Followed by auxiliary verbs ("do," "did," "does")

questions and "nor" clauses

No Comments

No Known Activations