INDEX

Explanations

words with the prefix "con-" followed by a noun or verb root

oai_token-act-pair · gpt-3.5-turbo

terms related to conformity and associated connotations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

baugh

-0.72

 calves

-0.64

 gelatin

-0.64

 footprints

-0.62

 Norn

-0.62

assetsadobe

-0.61

 acid

-0.61

 Waters

-0.61

lihood

-0.60

 shoots

-0.58

POSITIVE LOGITS

clave

1.27

clusively

1.26

undrum

1.24

gru

1.24

joined

1.19

notation

1.14

gres

1.13

naissance

1.09

gression

1.09

versely

1.09

Activations Density 0.026%

words with the prefix "con-" followed by a noun or verb root

terms related to conformity and associated connotations

No Comments

No Known Activations