INDEX

Explanations

quotes or insights from anonymous sources

oai_token-act-pair · gpt-3.5-turbo

phrases indicating sources of information or reports

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

Interstitial

-0.97

ãĤ¼ãĤ¦ãĤ¹

-0.77

Zen

-0.71

unicip

-0.71

enegger

-0.68

nect

-0.67

 Defeat

-0.67

wagen

-0.66

ennial

-0.65

fml

-0.65

POSITIVE LOGITS

 sources

0.89

ISI

0.84

pse

0.78

 Medline

0.77

 corrobor

0.74

 insiders

0.74

Ub

0.72

 informants

0.71

 knowledgeable

0.71

 investigator

0.71

Activations Density 0.160%

quotes or insights from anonymous sources

phrases indicating sources of information or reports

No Comments

No Known Activations