INDEX

Explanations

phrases that indicate purpose or intent

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

"for" followed by an article

np_acts-logits-general · gemini-2.0-flash

for your, for a, for direct, for Yugoslavia

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

 Cæsar

-0.66

 Euripides

-0.63

 Aphrodite

-0.62

 Shakspeare

-0.60

購入

-0.60

 createSprite

-0.58

pośred

-0.57

 cytok

-0.57

 uchiha

-0.56

POSITIVE LOGITS

the

1.28

0.94

 both

0.89

 those

0.87

geries

0.85

an

0.84

erun

0.83

 their

0.82

all

0.82

.}(

0.81

Activations Density 0.527%

phrases that indicate purpose or intent

"for" followed by an article

for your, for a, for direct, for Yugoslavia

No Comments

No Known Activations