INDEX

Explanations

references to academic institutions and individuals

oai_token-act-pair · gpt-3.5-turbo

names or terms associated with notable individuals and institutions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

iddles

-0.79

orthy

-0.75

apon

-0.75

ataka

-0.72

ocket

-0.69

Ń·

-0.69

ģĸ

-0.66

orsi

-0.66

olved

-0.65

 Misty

-0.65

POSITIVE LOGITS

stein

1.24

enegger

1.01

berg

0.93

erman

0.82

STE

0.81

 Fein

0.80

thal

0.79

itsch

0.77

 Feldman

0.76

enthal

0.75

Activations Density 0.026%

references to academic institutions and individuals

names or terms associated with notable individuals and institutions

No Comments

No Known Activations