INDEX

Explanations

phrases related to cooperation or conflict between groups or entities, particularly involving intelligence or political matters

oai_token-act-pair · gpt-3.5-turbo

references to individuals in a context of authority or influence

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

luster

-0.64

 corrid

-0.62

 Moroc

-0.62

 Barney

-0.62

 Clyde

-0.60

ADRA

-0.58

 advoc

-0.57

 Adin

-0.57

 Classics

-0.57

 Canal

-0.57

POSITIVE LOGITS

][

0.82

rael

0.73

 Reincarnated

0.73

"?

0.68

meet

0.66

ihad

0.66

>>\

0.65

soType

0.65

)?

0.65

Activations Density 0.355%

phrases related to cooperation or conflict between groups or entities, particularly involving intelligence or political matters

references to individuals in a context of authority or influence

No Comments

No Known Activations