INDEX

Explanations

verbs or phrases indicating potential threats or risks

oai_token-act-pair · gpt-3.5-turbo

phrases that describe potential threats

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

tery

-0.70

lex

-0.69

ocket

-0.68

Zup

-0.67

à¼

-0.62

 ILCS

-0.61

@#&

-0.60

 Nanto

-0.60

ergy

-0.60

Ago

-0.59

POSITIVE LOGITS

idon

1.16

 posed

0.89

atoon

0.83

 poses

0.82

 pose

0.81

 hazards

0.76

 dangers

0.72

 Danger

0.72

hran

0.71

 vulner

0.71

Activations Density 0.019%

verbs or phrases indicating potential threats or risks

phrases that describe potential threats

No Comments

No Known Activations