INDEX

Explanations

the word "a" followed by verbs or noun phrases indicating an action or entity

oai_token-act-pair · gpt-3.5-turbo

phrases that express the initiation of various new endeavors or projects

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

enance

-0.72

tips

-0.68

oyal

-0.64

itte

-0.64

cules

-0.64

Param

-0.63

tip

-0.62

do

-0.61

abytes

-0.61

Lens

-0.60

POSITIVE LOGITS

 fray

0.86

enium

0.73

 anew

0.72

 countdown

0.72

 journey

0.71

 conditioning

0.68

asus

0.67

ciating

0.67

 downward

0.66

 spiral

0.66

Activations Density 0.137%

the word "a" followed by verbs or noun phrases indicating an action or entity

phrases that express the initiation of various new endeavors or projects

No Comments

No Known Activations