INDEX

Explanations

phrases related to titles or headings with emphasis

oai_token-act-pair · gpt-3.5-turbo

words and phrases related to rules and agreements

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

etsk

-0.63

aday

-0.63

ko

-0.63

onym

-0.62

ciplinary

-0.61

ische

-0.60

virt

-0.59

cca

-0.59

taboola

-0.58

Ars

-0.58

POSITIVE LOGITS

URE

1.20

IES

1.20

LY

1.19

ING

1.14

URES

1.13

 WITH

1.12

ATED

1.12

DIT

1.12

ATIONS

1.12

ATES

1.10

Activations Density 0.280%

phrases related to titles or headings with emphasis

words and phrases related to rules and agreements

No Comments

No Known Activations