INDEX

Explanations

intensifiers that enhance the magnitude of negative evaluations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 0-gemmascope-mlp-65k

Configuration

google/gemma-scope-2b-pt-mlp/layer_0/width_65k/average_l0_72

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.0.hook_mlp_out

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

 incredibly

-0.79

 članak

-0.78

setopt

-0.76

 Roskov

-0.70

parsedMessage

-0.67

ThemeOverlay

-0.67

ArrowToggle

-0.67

 colpa

-0.66

 unbelievably

-0.66

gamesh

-0.66

POSITIVE LOGITS

ſelves

0.71

 Shakspeare

0.71

 Majefty

0.69

 Theſe

0.69

ſelf

0.68

 Monfieur

0.68

 myſelf

0.66

 ſeveral

0.66

_('

0.64

Vgl

0.63

Activations Density 0.013%

intensifiers that enhance the magnitude of negative evaluations

No Comments

No Known Activations