INDEX

Explanations

proper nouns or names containing 'H' followed by another high-activation character

oai_token-act-pair · gpt-3.5-turbo

words related to specific geographic locations and prominent individuals associated with them

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

taboola

-0.92

ãĥīãĥ©ãĤ´ãĥ³

-0.76

 Expend

-0.75

BILITIES

-0.71

 Cosponsors

-0.71

Ll

-0.70

BILITY

-0.65

Cause

-0.65

ioxide

-0.65

CLASSIFIED

-0.64

POSITIVE LOGITS

inki

1.00

iday

0.85

idan

0.85

wed

0.83

icultural

0.82

zik

0.81

kamp

0.78

itech

0.77

abee

0.76

awei

0.75

Activations Density 0.112%

proper nouns or names containing 'H' followed by another high-activation character

words related to specific geographic locations and prominent individuals associated with them

No Comments

No Known Activations