INDEX

Explanations

terms related to terrorist activity

oai_token-act-pair · gpt-3.5-turbo

instances of the letter 'j' 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 @ 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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ãĥ´ãĤ¡

-0.79

IUM

-0.69

 Carrier

-0.69

 Fargo

-0.66

 Scotia

-0.64

 DeVos

-0.63

 propensity

-0.63

IRA

-0.63

 Staples

-0.62

PORT

-0.62

POSITIVE LOGITS

igsaw

1.18

ingle

1.15

agged

1.13

ealous

1.12

ourn

1.09

upiter

1.09

umbled

1.09

aded

1.09

anky

1.08

itsu

1.08

Activations Density 0.016%

terms related to terrorist activity

instances of the letter 'j' in various contexts

No Comments

No Known Activations