INDEX

Explanations

phrases indicating comparisons or contrasts in various contexts

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

InjectAttribute

-0.74

AddTagHelper

-0.68

UserScript

-0.67

 ویکی‌پدی

-0.63

MessageTagHelper

-0.62

ReusableCell

-0.62

expandindo

-0.60

ConstraintMaker

-0.60

 houſe

-0.59

 Савезне

-0.56

POSITIVE LOGITS

legungen

0.37

 incó

0.34

rungsseite

0.33

 kır

0.32

 oyuncak

0.30

 måtte

0.30

 نوا

0.30

 inconsistent

0.29

 voet

0.29

 lediglich

0.29

Activations Density 0.070%

phrases indicating comparisons or contrasts in various contexts

No Comments

No Known Activations