INDEX

Explanations

phrases or words related to authoritative or powerful actions

oai_token-act-pair · gpt-3.5-turbo

instances of the letter "v" 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 @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 Mellon

-0.71

ĪĴ

-0.70

Ĥª

-0.67

 CLSID

-0.67

 Chaser

-0.65

ptr

-0.65

Ķ

-0.61

 McKay

-0.61

 Rasmussen

-0.60

ħĭ

-0.60

POSITIVE LOGITS

ulnerable

1.42

ortex

1.41

iral

1.39

irgin

1.39

apor

1.38

igil

1.38

ulnerability

1.36

ibr

1.35

ampires

1.35

isions

1.33

Activations Density 0.033%

phrases or words related to authoritative or powerful actions

instances of the letter "v" in various contexts

No Comments

No Known Activations