INDEX

Explanations

mentions of individuals or entities being championed or recognized as champions

oai_token-act-pair · gpt-3.5-turbo

instances of the word "champion" in various contexts

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

ortunate

-0.76

kefeller

-0.75

eret

-0.75

fits

-0.73

uld

-0.73

iple

-0.73

plet

-0.72

yden

-0.69

ali

-0.69

sis

-0.68

POSITIVE LOGITS

 champion

1.30

 champions

1.22

 Champion

1.13

ampion

0.93

ampions

0.84

 challenger

0.80

naire

0.80

 champ

0.79

 Chall

0.77

 boxer

0.76

Activations Density 0.010%

mentions of individuals or entities being championed or recognized as champions

instances of the word "champion" in various contexts

No Comments

No Known Activations