INDEX

Explanations

names and terms related to individuals and organizations

oai_token-act-pair · gpt-3.5-turbo

references to specific names and terms associated with individuals and organizations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

earchers

-1.03

areth

-0.78

eah

-0.77

aped

-0.77

ateurs

-0.74

asury

-0.72

phis

-0.71

abetic

-0.70

ierre

-0.70

seys

-0.70

POSITIVE LOGITS

Boh

0.92

boarding

0.85

SourceFile

0.76

rences

0.73

flo

0.72

gettable

0.70

 Wick

0.70

 Wink

0.70

 Airl

0.69

Brow

0.68

Activations Density 0.029%

names and terms related to individuals and organizations

references to specific names and terms associated with individuals and organizations

No Comments

No Known Activations