INDEX

Explanations

words related to places, names, and titles

oai_token-act-pair · gpt-3.5-turbo

words related to names and names of people

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

glers

-0.76

 blunt

-0.61

 academ

-0.61

irlf

-0.58

ickr

-0.57

annels

-0.57

anchester

-0.57

 compe

-0.56

 magnification

-0.56

 slash

-0.56

POSITIVE LOGITS

ette

0.81

atis

0.77

ettes

0.76

nih

0.70

idge

0.69

enne

0.68

akis

0.67

agos

0.65

seys

0.65

Royal

0.64

Activations Density 0.058%

words related to places, names, and titles

words related to names and names of people

No Comments

No Known Activations