INDEX

Explanations

names of individuals

oai_token-act-pair · gpt-3.5-turbo

proper nouns, specifically names of individuals

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

 includ

-0.74

ccording

-0.72

 looph

-0.68

 NETWORK

-0.67

EStream

-0.64

 suspic

-0.64

 destro

-0.63

 tiss

-0.62

ortium

-0.61

benefit

-0.59

POSITIVE LOGITS

 alike

1.05

 respectively

0.78

axter

0.75

 versa

0.75

oliath

0.73

VB

0.67

Ru

0.64

ilda

0.63

avia

0.62

 thereof

0.62

Activations Density 0.361%

names of individuals

proper nouns, specifically names of individuals

No Comments

No Known Activations