INDEX

Explanations

words related to personal or moral attributes

oai_token-act-pair · gpt-3.5-turbo

references to "character" in various contexts

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

IDA

-0.73

yg

-0.70

hub

-0.68

EMS

-0.63

algia

-0.63

hner

-0.62

xon

-0.62

yah

-0.61

yden

-0.61

ESP

-0.60

POSITIVE LOGITS

istics

1.44

istically

1.41

izations

1.18

izes

1.07

isations

0.97

acters

0.97

 assassination

0.87

isation

0.86

 charact

0.86

ised

0.85

Activations Density 0.022%

words related to personal or moral attributes

references to "character" in various contexts

No Comments

No Known Activations