INDEX

Explanations

phrases related to control or authority

oai_token-act-pair · gpt-3.5-turbo

phrases emphasizing the concept of allowing or permitting actions or behaviors

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

oppable

-0.83

 Languages

-0.72

 holiest

-0.69

lihood

-0.67

atana

-0.67

 millenn

-0.66

agher

-0.66

querque

-0.63

 cumbers

-0.63

edly

-0.63

POSITIVE LOGITS

tered

1.12

tering

0.92

 slip

0.87

icia

0.86

enne

0.75

 loose

0.75

us

0.73

itia

0.69

go

0.67

me

0.67

Activations Density 0.028%

phrases related to control or authority

phrases emphasizing the concept of allowing or permitting actions or behaviors

No Comments

No Known Activations