INDEX

Explanations

instances of the word "won't" in various forms

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

New Auto-Interp

Configuration

google/gemma-scope-2b-pt-res/layer_12/width_1m/average_l0_107

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

1,048,576

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

steuer

-0.75

 Humphreys

-0.75

 squares

-0.68

Laure

-0.67

 campana

-0.66

 Squares

-0.66

 grø

-0.66

NoSuchAlgorithm

-0.65

 nuages

-0.65

 Newman

-0.65

POSITIVE LOGITS

 wont

1.14

 WILL

0.97

不会

0.81

Willy

0.81

expandindo

0.80

 Willy

0.79

 Gonna

0.79

WILL

0.78

 Will

0.78

 will

0.75

Activations Density 0.000%

instances of the word "won't" in various forms

No Comments

No Known Activations