INDEX

Explanations

negatives emphasizing the quality or nature of something being unfavorable

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

ConstraintMaker

-0.55

 NSCoder

-0.55

 ivelany

-0.55

 duquel

-0.53

absolutely

-0.51

しまいます

-0.50

しまう

-0.49

finally

-0.49

Boring

-0.49

ってしまう

-0.48

POSITIVE LOGITS

 pleasant

0.92

 good

0.81

 conducive

0.72

 bueno

0.71

 welcome

0.71

 desirable

0.71

 nice

0.69

 ideal

0.68

 favorable

0.67

 bode

0.66

Activations Density 0.255%

negatives emphasizing the quality or nature of something being unfavorable

No Comments

No Known Activations