INDEX

Explanations

patterns of visual shapes in a structured layout

oai_token-act-pair · gpt-3.5-turbo

characters and symbols in complex formats, likely related to visual or spatial representations

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

glers

-0.42

emort

-0.36

éĹĺ

-0.36

aution

-0.34

iaries

-0.34

 Gutenberg

-0.34

okÃ©

-0.33

 convened

-0.32

allery

-0.32

 conduc

-0.32

POSITIVE LOGITS

ax

0.37

lite

0.33

resp

0.33

parent

0.33

bt

0.33

))

0.32

")

0.32

\\

0.32

cast

0.31

ÙĨ

0.31

Activations Density 1.340%

patterns of visual shapes in a structured layout

characters and symbols in complex formats, likely related to visual or spatial representations

No Comments

No Known Activations