INDEX

Explanations

proper nouns

oai_token-act-pair · gpt-3.5-turbo

mentions of the name "Kra" and variations, along with the name "Kab."

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

payer

-0.86

orative

-0.81

sheet

-0.76

 Beckham

-0.72

hips

-0.71

iants

-0.71

-0.70

 sheets

-0.70

 à¨

-0.69

rics

-0.69

POSITIVE LOGITS

Kra

1.19

ven

0.92

uth

0.88

Å¡

0.87

emer

0.83

lyak

0.83

unte

0.82

plin

0.81

 destro

0.80

assic

0.78

Activations Density 0.005%

proper nouns

mentions of the name "Kra" and variations, along with the name "Kab."

No Comments

No Known Activations