INDEX

Explanations

references to exhibitions or displays

oai_token-act-pair · gpt-3.5-turbo

instances of the word "exhibit" 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

pop

-0.75

sembly

-0.74

tarian

-0.74

yip

-0.73

leans

-0.69

shaw

-0.69

kos

-0.68

 Wast

-0.66

zbek

-0.65

nian

-0.65

POSITIVE LOGITS

ĸļ

1.03

 exhibits

0.98

 exhibit

0.90

âĸº

0.78

 exhib

0.78

emonium

0.77

 Exhibit

0.77

hibit

0.77

ttes

0.77

ivities

0.77

Activations Density 0.011%

references to exhibitions or displays

instances of the word "exhibit" and its variations

No Comments

No Known Activations