INDEX

Explanations

phrases or words with the structure "in 'something'"

oai_token-act-pair · gpt-3.5-turbo

instances of the word "in" and variations of phrases involving locations or contexts

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

ADRA

-0.83

 vanquished

-0.80

 Seym

-0.75

 attackers

-0.67

 Mour

-0.66

anasia

-0.66

 allerg

-0.64

 Shutterstock

-0.64

 Gutenberg

-0.64

witz

-0.64

POSITIVE LOGITS

tha

0.90

appropriate

0.86

 gonna

0.82

nit

0.75

wards

0.72

pace

0.70

selves

0.70

cause

0.70

\'

0.69

ta

0.69

Activations Density 0.064%

phrases or words with the structure "in 'something'"

instances of the word "in" and variations of phrases involving locations or contexts

No Comments

No Known Activations