INDEX

Explanations

capitalized verbs related to actions or initiatives

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of the words "Take," "Give," "Save," and "Leave" related to campaigns or calls to action

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

 exha

-0.82

 destro

-0.72

opausal

-0.70

 discharge

-0.70

 tremend

-0.69

etheless

-0.67

 withd

-0.67

 toget

-0.65

)].

-0.64

 constitu

-0.64

POSITIVE LOGITS

ings

1.07

away

1.03

aways

0.92

ables

0.92

ners

0.88

 Yourself

0.88

 Your

0.87

lihood

0.86

ership

0.85

ers

0.84

Activations Density 0.199%

capitalized verbs related to actions or initiatives

repeated mentions of the words "Take," "Give," "Save," and "Leave" related to campaigns or calls to action

No Comments

No Known Activations