INDEX

Explanations

phrases related to comparison or belonging

oai_token-act-pair · gpt-3.5-turbo

phrases related to majority or significant numbers

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

 referen

-0.61

 Seym

-0.53

 Moroc

-0.52

 Vaugh

-0.52

 thous

-0.50

chuk

-0.49

edIn

-0.48

ãĥĩãĤ£

-0.47

 helicop

-0.45

 corrid

-0.45

POSITIVE LOGITS

 largeDownload

0.58

othes

0.47

mosp

0.46

 DRAGON

0.45

immer

0.45

taboola

0.45

ciples

0.44

cing

0.44

rg

0.43

hard

0.43

Activations Density 3.359%

phrases related to comparison or belonging

phrases related to majority or significant numbers

No Comments

No Known Activations