INDEX

Explanations

words related to headlines or titles, such as names and dates

oai_token-act-pair · gpt-3.5-turbo

instances of the letter 'S' in various contexts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Strip

-0.68

CPR

-0.68

 Ribbon

-0.66

 occupied

-0.64

 psycho

-0.61

 Proto

-0.61

 formerly

-0.60

 Kali

-0.59

 stressed

-0.59

 unrestricted

-0.59

POSITIVE LOGITS

ixty

1.38

atisf

1.34

ometime

1.33

omew

1.33

olutions

1.33

ierra

1.30

atellite

1.29

avage

1.27

essions

1.25

outheast

1.23

Activations Density 0.039%

words related to headlines or titles, such as names and dates

instances of the letter 'S' in various contexts

No Comments

No Known Activations