INDEX

Explanations

words related to accusation, blame, or criticism

oai_token-act-pair · gpt-3.5-turbo

terms related to appropriation and scapegoating

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

hands

-0.71

QUIRE

-0.66

 Madison

-0.66

locks

-0.65

 Driver

-0.65

catentry

-0.65

Ov

-0.65

mith

-0.64

terday

-0.64

DEP

-0.64

POSITIVE LOGITS

Ĥİ

0.81

anism

0.81

eness

0.79

issance

0.78

 Sabha

0.78

asting

0.76

ating

0.73

asted

0.73

vil

0.72

wikipedia

0.72

Activations Density 0.033%

words related to accusation, blame, or criticism

terms related to appropriation and scapegoating

No Comments

No Known Activations