INDEX

Explanations

words related to weakness or vulnerability

oai_token-act-pair · gpt-3.5-turbo

references to weakness or frailty

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

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

ICAN

-0.81

APH

-0.75

agher

-0.74

andise

-0.71

alogue

-0.71

 Noir

-0.69

 Sloan

-0.68

 Andromeda

-0.68

 McCann

-0.68

ICA

-0.67

POSITIVE LOGITS

nesses

1.27

lings

1.18

ling

0.99

ening

0.92

ens

0.91

ener

0.87

ened

0.86

est

0.85

les

0.84

 minded

0.82

Activations Density 0.013%

words related to weakness or vulnerability

references to weakness or frailty

No Comments

No Known Activations