INDEX

Explanations

interjections such as "So" used to introduce thoughts or explanations

oai_token-act-pair · gpt-3.5-turbo

repeated phrases that start with "So."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

 presentation

-0.66

idity

-0.64

``(

-0.63

.","

-0.62

ãĤ¼ãĤ¦ãĤ¹

-0.60

äºº

-0.60

 Peninsula

-0.60

"],"

-0.59

enment

-0.58

''

-0.58

POSITIVE LOGITS

ooo

1.00

oooo

0.97

othes

0.94

oths

0.93

bered

0.92

oner

0.91

ppy

0.90

arin

0.89

fter

0.89

aked

0.87

Activations Density 0.055%

interjections such as "So" used to introduce thoughts or explanations

repeated phrases that start with "So."

No Comments

No Known Activations