INDEX

Explanations

named entities

oai_token-act-pair · gpt-3.5-turbo

proper nouns, particularly names and entities

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

hement

-0.64

 compr

-0.64

alus

-0.60

 corpor

-0.60

âķĲâķĲ

-0.58

ãĢĲ

-0.57

mosp

-0.57

ghai

-0.56

 ..........

-0.54

arma

-0.54

POSITIVE LOGITS

ECK

0.62

KE

0.62

EEK

0.60

 Voting

0.58

xit

0.57

DAY

0.57

STON

0.56

 quotes

0.55

 Jinn

0.54

GO

0.54

Activations Density 1.146%

named entities

proper nouns, particularly names and entities

No Comments

No Known Activations