INDEX

Explanations

strong modifiers indicating positivity or emphasis

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

New Auto-Interp

Configuration

google/gemma-scope-2b-pt-res/layer_12/width_32k/average_l0_76

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

32,768

Data Type

float32

Hook Name

blocks.12.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

upaten

-0.73

Datuak

-0.69

 oblivion

-0.69

 Miyazaki

-0.68

 bliss

-0.68

 Ghosts

-0.67

 calendars

-0.66

XmlAccessorType

-0.66

Hozzáférés

-0.66

 Maus

-0.66

POSITIVE LOGITS

 Strong

1.47

strong

1.45

STRONG

1.42

 STRONG

1.41

Strong

1.41

 strong

1.38

ngths

1.02

strength

1.02

 stronger

1.00

 Stronger

0.97

Activations Density 0.039%

strong modifiers indicating positivity or emphasis

No Comments

No Known Activations