INDEX

Explanations

phrases indicating consideration or comparison

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

eday

-0.57

entin

-0.52

Dialogue

-0.51

eches

-0.51

enger

-0.49

ezvous

-0.47

 Reserved

-0.46

assad

-0.46

paren

-0.45

 battered

-0.45

POSITIVE LOGITS

Wr

0.63

opsis

0.55

othy

0.52

tails

0.51

meaning

0.51

false

0.50

eye

0.50

oa

0.50

ovi

0.49

SourceFile

0.49

Activations Density 0.255%

phrases indicating consideration or comparison

No Comments

No Known Activations