INDEX

Explanations

declarations or statements made with conviction or authority

oai_token-act-pair · gpt-3.5-turbo

instances of the word "declared" and its variations

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

 captcha

-0.72

umbn

-0.68

Reply

-0.68

nesota

-0.68

ROR

-0.66

committee

-0.65

«ĺ

-0.62

dayName

-0.62

ĸļ

-0.61

alez

-0.61

POSITIVE LOGITS

 bankruptcy

0.89

ignty

0.83

 allegiance

0.78

phas

0.78

 declares

0.77

 unequivocally

0.76

arations

0.75

 declaring

0.73

unci

0.73

 martial

0.73

Activations Density 0.023%

declarations or statements made with conviction or authority

instances of the word "declared" and its variations

No Comments

No Known Activations