INDEX

Explanations

words related to fires and arson

oai_token-act-pair · gpt-3.5-turbo

references to fire and arson-related incidents

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

allery

-0.73

 Surgery

-0.70

atem

-0.67

ajor

-0.66

Vag

-0.65

 Senators

-0.65

apore

-0.65

glomer

-0.64

omo

-0.61

afort

-0.60

POSITIVE LOGITS

 blaze

1.31

flame

0.98

flies

0.97

 blazing

0.94

 fires

0.93

 flames

0.93

 extingu

0.90

 exting

0.89

 extinguished

0.88

lda

0.85

Activations Density 0.017%

words related to fires and arson

references to fire and arson-related incidents

No Comments

No Known Activations