INDEX

Explanations

the word "do" and phrases indicating importance or significance in context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 4-gemmascope-mlp-65k

Configuration

google/gemma-scope-2b-pt-mlp/layer_4/width_65k/average_l0_66

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.4.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

olesome

-0.50

 carry

-0.50

>[

-0.47



-0.47

โม

-0.45

ządu

-0.45

forbes

-0.45

anskje

-0.44

feira

-0.44

OfWork

-0.44

POSITIVE LOGITS

GraphicsUnit

0.78

 propOrder

0.76

 GenerationType

0.75

InvalidProtocol

0.73

 nahilalakip

0.72

 myſelf

0.71

Efq

0.68

BufferException

0.67

ReusableCell

0.67

BeginContext

0.67

Activations Density 0.006%

the word "do" and phrases indicating importance or significance in context

No Comments

No Known Activations