INDEX

Explanations

the specific term "Ass" with varying activation strengths

oai_token-act-pair · gpt-3.5-turbo

variations of the term "assassin" or related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

çĦ

-0.72

AAF

-0.69

vation

-0.68

 Tobacco

-0.64

 Mercury

-0.63

WD

-0.63

 Welsh

-0.63

 poppy

-0.62

MET

-0.62

 smoke

-0.61

POSITIVE LOGITS

assin

1.36

ociate

1.28

essor

1.27

sembly

1.21

ass

1.21

alam

1.16

oci

1.13

ortment

1.12

alon

1.09

essment

1.08

Activations Density 0.007%

the specific term "Ass" with varying activation strengths

variations of the term "assassin" or related terms

No Comments

No Known Activations