INDEX

Explanations

phrases related to judgments or opinions

oai_token-act-pair · gpt-3.5-turbo

expressions of judgment and significant actions

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

•Voight-Kampff

No Comments

Negative Logits

erity

-0.59

ensibly

-0.57

nesday

-0.56

enta

-0.56

oaded

-0.52

then

-0.52

ensing

-0.51

erence

-0.51

sts

-0.51

reements

-0.50

POSITIVE LOGITS

 urban

0.59

 doping

0.55

 recreational

0.55

 culinary

0.55

 international

0.54

 academic

0.52

 sports

0.52

 tourist

0.51

 artistic

0.51

 commercial

0.51

Activations Density 1.761%

phrases related to judgments or opinions

expressions of judgment and significant actions

No Comments

No Known Activations