INDEX

Explanations

numerical sequences and technical jargon

oai_token-act-pair · gpt-3.5-turbo

structured information about geographical locations and statistics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

Offline

-0.73

espie

-0.72

Reviewed

-0.70

taboola

-0.64

advertising

-0.62

interstitial

-0.59

 magnification

-0.59

entin

-0.57

wagen

-0.57

lde

-0.56

POSITIVE LOGITS

ucc

0.64

llah

0.63

 Beckham

0.63

querque

0.60

 Pearl

0.56

 Samar

0.54

ICAN

0.53

ufact

0.53

 Shak

0.50

arah

0.50

Activations Density 1.363%

numerical sequences and technical jargon

structured information about geographical locations and statistics

No Comments

No Known Activations