INDEX

Explanations

words related to exuding strength or impact, such as 'impresses' or 'affecting'

oai_token-act-pair · gpt-3.5-turbo

statistical data or numerical information related to various topics

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

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

icum

-0.63

ilic

-0.57

 Dragonbound

-0.53

ently

-0.52

WATCH

-0.52

oppy

-0.50

otropic

-0.50

isp

-0.50

pta

-0.48

AMA

-0.48

POSITIVE LOGITS

 from

1.82

from

1.73

 FROM

1.59

From

1.52

 From

1.51

fro

0.83

 originating

0.69

 through

0.67

 derived

0.66

 thence

0.66

Activations Density 0.257%

words related to exuding strength or impact, such as 'impresses' or 'affecting'

statistical data or numerical information related to various topics

No Comments

No Known Activations