INDEX

Explanations

phrases related to fictional characters or organizations

oai_token-act-pair · gpt-3.5-turbo

references to the word "ami."

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

keye

-0.80

sed

-0.75

 Vaugh

-0.69

sie

-0.65

ription

-0.65

tain

-0.62

rants

-0.62

drive

-0.62

tarians

-0.61

keyes

-0.61

POSITIVE LOGITS

ibo

1.27

yah

1.01

Äĩ

0.98

isance

0.97

emi

0.93

zzle

0.92

Ã±a

0.92

BILITY

0.83

roth

0.83

pta

0.83

Activations Density 0.013%

phrases related to fictional characters or organizations

references to the word "ami."

No Comments

No Known Activations