INDEX

Explanations

phrases indicating going back in time or reverting to a previous state or action

oai_token-act-pair · gpt-3.5-turbo

phrases that reference returning to previous times or states

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

merce

-0.91

ially

-0.78

 Donation

-0.75

é¾įåĸļå£«

-0.75

ello

-0.74

croft

-0.72

tesy

-0.69

stals

-0.69

riad

-0.68

owler

-0.68

POSITIVE LOGITS

 haunt

1.10

 whence

0.89

 shore

0.83

 basics

0.77

 sanity

0.73

forth

0.73

 earth

0.69

 grid

0.68

 forth

0.67

 civilisation

0.66

Activations Density 0.062%

phrases indicating going back in time or reverting to a previous state or action

phrases that reference returning to previous times or states

No Comments

No Known Activations