INDEX

Explanations

phrases indicating negative outcomes or challenges

oai_token-act-pair · gpt-3.5-turbo

negative developments or deteriorating situations

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ngth

-0.91

uni

-0.69

thood

-0.68

gemony

-0.67

 superior

-0.66

otype

-0.66

ools

-0.65

license

-0.65

iciency

-0.65

pine

-0.64

POSITIVE LOGITS

 downhill

0.77

:#

0.69

 hinges

0.69

 unfold

0.68

 Thrones

0.66

oulos

0.64

 suspense

0.64

 Manafort

0.64

EMS

0.64

 Yanukovych

0.63

Activations Density 0.422%

phrases indicating negative outcomes or challenges

negative developments or deteriorating situations

No Comments

No Known Activations