INDEX

Explanations

HTML tags and attributes, particularly metadata and titles

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

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

 named

-0.97

 name

-0.94

 Name

-0.92

Name

-0.87

 Namen

-0.87

name

-0.86

 Named

-0.86

 naming

-0.84

Named

-0.80

 nommé

-0.77

POSITIVE LOGITS

 subtitle

0.77

Subtitle

0.72

 Subtitle

0.66

 description

0.65

description

0.63

subtitle

0.62

副

0.60

 SafeMath

0.60

adaptiveStyles

0.58

getDescription

0.56

Activations Density 0.207%

HTML tags and attributes, particularly metadata and titles

No Comments

No Known Activations