INDEX

Explanations

words related to causality or explanation

oai_token-act-pair · gpt-3.5-turbo

instances of the word "hence" and its variations

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

 rehabilit

-0.68

 hitter

-0.64

abies

-0.60

 nurs

-0.56

 batter

-0.56

 territ

-0.55

 rubbish

-0.54

igger

-0.53

 neighb

-0.53

 dentist

-0.53

POSITIVE LOGITS

forth

2.03

forward

1.38

why

0.98

why

0.87

xual

0.86

entimes

0.84

noon

0.77

fter

0.76

videos

0.75

ãĤ´ãĥ³

0.73

Activations Density 0.034%

words related to causality or explanation

instances of the word "hence" and its variations

No Comments

No Known Activations