INDEX

Explanations

submissions and references labeled with "Sub" followed by a numerical code

oai_token-act-pair · gpt-3.5-turbo

the presence of the word "Sub" used in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Cors

-0.74

ILLE

-0.73

 Zurich

-0.73

rot

-0.71

 Jenner

-0.68

 Muse

-0.67

 Papers

-0.66

OHN

-0.65

fet

-0.65

fir

-0.64

POSITIVE LOGITS

sequent

1.62

mitted

1.58

sequently

1.57

stantial

1.52

mitting

1.33

scription

1.32

ordinate

1.31

verting

1.30

missive

1.30

stant

1.29

Activations Density 0.021%

submissions and references labeled with "Sub" followed by a numerical code

the presence of the word "Sub" used in various contexts

No Comments

No Known Activations