INDEX

Explanations

references to the character "Holmes" with varying degrees of activation intensity

oai_token-act-pair · gpt-3.5-turbo

references to characters or elements from the Sherlock Holmes stories

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

ital

-0.86

nia

-0.79

icas

-0.78

ITIES

-0.77

illian

-0.74

arian

-0.74

acious

-0.71

itative

-0.71

imens

-0.70

ities

-0.70

POSITIVE LOGITS

 Flavoring

1.01

mble

0.94

ĸļ

0.89

ecake

0.80

ï¸

0.77

lling

0.68

´

0.67

xual

0.66

ĪĴ

0.66

llers

0.64

Activations Density 0.058%

references to the character "Holmes" with varying degrees of activation intensity

references to characters or elements from the Sherlock Holmes stories

No Comments

No Known Activations