INDEX

Explanations

keywords related to seeking information and asking questions starting with "How"

oai_token-act-pair · gpt-3.5-turbo

questions beginning with "How."

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

Ãį

-0.62

iculture

-0.61

 fide

-0.59

 receiving

-0.58

yak

-0.57

outer

-0.57

agonists

-0.56

uthor

-0.56

IENT

-0.56

ceptions

-0.55

POSITIVE LOGITS

soever

1.16

ever

0.99

beit

0.96

itzer

0.81

ling

0.80

bill

0.79

ells

0.78

leep

0.78

ls

0.73

lers

0.71

Activations Density 0.067%

keywords related to seeking information and asking questions starting with "How"

questions beginning with "How."

No Comments

No Known Activations