INDEX

Explanations

the word "Jerusalem"

oai_token-act-pair · gpt-3.5-turbo

mentions of Jerusalem

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

emonium

-0.76

ramid

-0.74

lass

-0.74

ority

-0.72

ãģ¦

-0.71

ured

-0.71

urer

-0.70

 Jackets

-0.70

raviolet

-0.70

rha

-0.69

POSITIVE LOGITS

usalem

1.12

 Jerusalem

0.95

 Pradesh

0.82

 Municip

0.81

ites

0.78

 municipality

0.75

 Aviv

0.73

itable

0.70

 correspondent

0.70

 Township

0.68

Activations Density 0.013%

the word "Jerusalem"

mentions of Jerusalem

No Known Activations