INDEX

Explanations

terms related to powerful entities or organizations, particularly those associated with influence and control

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

*/(

-0.74

Acknowled

-0.73

Quality

-0.72

yrics

-0.69

 Quality

-0.68

etooth

-0.67

Stand

-0.67

Ada

-0.67

uctor

-0.66

Board

-0.65

POSITIVE LOGITS

 empire

1.16

 empires

0.97

ipel

0.92

cles

0.89

 abroad

0.83

 colon

0.77

 hegemony

0.76

 plunder

0.75

 conquered

0.75

urious

0.74

Activations Density 0.044%

terms related to powerful entities or organizations, particularly those associated with influence and control

No Comments

No Known Activations