INDEX

Explanations

phrases related to breaking or disruption

oai_token-act-pair · gpt-3.5-turbo

references to a specific television show or media franchise

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ser

-0.75

minist

-0.70

sal

-0.67

ient

-0.66

 administrative

-0.65

ur

-0.64

 potent

-0.64

 latitude

-0.62

hed

-0.62

 fearful

-0.62

POSITIVE LOGITS

 Break

3.81

Break

2.50

break

1.83

 break

1.58

 Breaker

1.47

 breakdown

1.31

 breaks

1.30

 Broken

1.29

 Breaking

1.29

 Blow

1.25

Activations Density 0.017%

phrases related to breaking or disruption

references to a specific television show or media franchise

No Comments

No Known Activations