INDEX

Explanations

phrases related to positivity or benefit

oai_token-act-pair · gpt-3.5-turbo

phrases conveying positive assessments or conditions

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

RTX

-0.59

gall

-0.56

pa

-0.55

cheat

-0.55

egu

-0.55

hao

-0.54

pipe

-0.54

pred

-0.53

olate

-0.53

bil

-0.53

POSITIVE LOGITS

 sake

1.52

 reasons

1.26

ummies

1.23

 purposes

1.18

 reason

1.01

 Reasons

0.86

 ages

0.78

aughs

0.76

ankind

0.75

teness

0.73

Activations Density 0.173%

phrases related to positivity or benefit

phrases conveying positive assessments or conditions

No Comments

No Known Activations