INDEX

Explanations

phrases related to acquiring knowledge or information

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the act of learning or receiving news

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 exting

-0.73

 teasp

-0.68

enture

-0.66

ascript

-0.65

 privile

-0.64

pmwiki

-0.63

paio

-0.62

oreAnd

-0.62

ĸļ

-0.61

respective

-0.61

POSITIVE LOGITS

 that

0.76

 about

0.70

how

0.70

 goodbye

0.70

 DeVos

0.68

he

0.68

 news

0.67

she

0.67

 what

0.66

noon

0.64

Activations Density 0.126%

phrases related to acquiring knowledge or information

phrases indicating the act of learning or receiving news

No Comments

No Known Activations