INDEX

Explanations

words related to power or leadership

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

 shortened

-0.64

ãĤ´ãĥ³

-0.60

 resolving

-0.59

ãĤ´

-0.58

 Escape

-0.58

 Chaser

-0.58

 manners

-0.58

 commissions

-0.57

 Trib

-0.57

 evenings

-0.56

POSITIVE LOGITS

arsity

1.45

oodoo

1.43

anguard

1.37

ulture

1.34

iking

1.24

alkyrie

1.19

incent

1.16

andals

1.14

itals

1.13

ancouver

1.11

Activations Density 0.025%

words related to power or leadership

No Comments

No Known Activations