INDEX

Explanations

words related to unity or joining together

oai_token-act-pair · gpt-3.5-turbo

variations of the root word "unite" and its derivatives

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

dod

-0.68

 delinqu

-0.68

cale

-0.67

vag

-0.64

deb

-0.63

eah

-0.63

 inacc

-0.63

espie

-0.62

efully

-0.60

expensive

-0.60

POSITIVE LOGITS

States

0.72

 Kingdoms

0.72

 Realms

0.71

oppable

0.69

 footing

0.69

 territories

0.68

atives

0.68

 Nations

0.68

arity

0.68

owship

0.67

Activations Density 0.039%

words related to unity or joining together

variations of the root word "unite" and its derivatives

No Comments

No Known Activations