INDEX

Explanations

phrases related to accountability and responsibility

oai_token-act-pair · gpt-3.5-turbo

instances of the word "hold" and its variations 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 @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

————

-0.70

ãĤ¡

-0.69

ghan

-0.67

endix

-0.66

zyk

-0.65

ibel

-0.64

ettel

-0.64

gnu

-0.64

shire

-0.63

nown

-0.62

POSITIVE LOGITS

 sway

1.06

hold

1.01

erness

1.01

holders

0.95

holding

0.94

 accountable

0.94

 onto

0.92

overs

0.89

holder

0.89

fast

0.83

Activations Density 0.058%

phrases related to accountability and responsibility

instances of the word "hold" and its variations in various contexts

No Comments

No Known Activations