INDEX

Explanations

phrases indicating a sequence of actions or steps

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate a sequence of actions or requirements

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

milo

-0.68

tex

-0.67

tek

-0.67

spl

-0.66

tur

-0.66

haired

-0.65

dden

-0.65

rower

-0.64

tu

-0.63

reath

-0.62

POSITIVE LOGITS

Osw

0.88

ãĥĻ

0.75

lies

0.75

liness

0.71

ppel

0.67

vity

0.67

 fulfillment

0.66

 coer

0.65

zzo

0.65

arity

0.64

Activations Density 0.019%

phrases indicating a sequence of actions or steps

phrases that indicate a sequence of actions or requirements

No Comments

No Known Activations