INDEX

Explanations

phrases related to propaganda

oai_token-act-pair · gpt-3.5-turbo

references to specific propositions or legislation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

keeper

-0.65

 Caribbean

-0.64

 horizont

-0.62

cence

-0.61

gently

-0.61

 charism

-0.60

 pores

-0.59

Maw

-0.58

 Peaks

-0.58

 blat

-0.58

POSITIVE LOGITS

aganda

1.62

ulsion

1.42

onents

1.32

osed

1.31

ulsive

1.28

itious

1.25

ylene

1.23

onent

1.22

ensity

1.18

osition

1.16

Activations Density 0.023%

phrases related to propaganda

references to specific propositions or legislation

No Comments

No Known Activations