INDEX

Explanations

the word "hell" and its variations, often in emotional or negative contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 5-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_5/width_16k/average_l0_68

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.5.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

vábbi

-0.78

AndEndTag

-0.77

rungsseite

-0.74

subpackage

-0.73

 Bisch

-0.72

Démographie

-0.72

 Bourgoin

-0.72

 VIAF

-0.71

rijke

-0.70

ῖν

-0.68

POSITIVE LOGITS

 hell

1.97

 Hell

1.95

Hell

1.77

 HELL

1.74

hell

1.35

HELL

1.26

 infierno

1.26

 heck

1.16

Heck

1.05

 Heck

0.99

Activations Density 0.060%

the word "hell" and its variations, often in emotional or negative contexts

No Comments

No Known Activations