INDEX

Explanations

names or terms related to locations or people

oai_token-act-pair · gpt-3.5-turbo

names of people and places

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ãĥĺãĥ©

-0.96

 Imag

-0.86

transform

-0.85

 Dynam

-0.79

ãĤ¨

-0.78

 Glow

-0.78

iP

-0.78

 istg

-0.76

èĥ

-0.75

ANG

-0.74

POSITIVE LOGITS

ley

1.27

leys

1.01

roy

0.96

sey

0.94

lay

0.94

LEY

0.93

ameron

0.90

ey

0.90

ay

0.89

aryl

0.86

Activations Density 0.258%

names or terms related to locations or people

names of people and places

No Comments

No Known Activations