INDEX

Explanations

proper nouns or specific names

oai_token-act-pair · gpt-3.5-turbo

phrases related to the title and content of specific television series or films

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

usercontent

-0.62

 bases

-0.58

ciplinary

-0.56

ĸļ

-0.56

chnology

-0.54

soDeliveryDate

-0.53

 probes

-0.53

 surveys

-0.53

channelAvailability

-0.52

alore

-0.52

POSITIVE LOGITS

âĻ¥

0.57

uly

0.56

 Meditation

0.53

iva

0.53

stic

0.52

gp

0.51

 outp

0.51

heit

0.51

ili

0.50

Ń·

0.50

Activations Density 0.889%

proper nouns or specific names

phrases related to the title and content of specific television series or films

No Comments

No Known Activations