INDEX

Explanations

phrases related to censorship and suppression

oai_token-act-pair · gpt-3.5-turbo

actions related to suppression and censorship

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

ortment

-0.78

esides

-0.65

luster

-0.65

Suc

-0.65

immer

-0.64

ammy

-0.64

olds

-0.64

âĹ¼

-0.63

-------

-0.61

itched

-0.61

POSITIVE LOGITS

ively

0.92

uate

0.81

him

0.79

 offending

0.78

 opposing

0.77

enance

0.75

 them

0.73

 everything

0.72

 dissent

0.72

 oneself

0.71

Activations Density 0.199%

phrases related to censorship and suppression

actions related to suppression and censorship

No Comments

No Known Activations