INDEX

Explanations

names of specific individuals, potentially related to sports or political figures

oai_token-act-pair · gpt-3.5-turbo

proper nouns, particularly 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 @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

inel

-0.77

alogue

-0.77

isl

-0.76

oved

-0.76

than

-0.73

atan

-0.73

 unden

-0.73

ern

-0.71

evil

-0.71

inet

-0.66

POSITIVE LOGITS

 Fitzpatrick

0.93

patrick

0.85

 Alley

0.82

hler

0.74

eru

0.73

 Ingram

0.72

 Pryor

0.72

 Cummings

0.68

 Dickinson

0.68

 Shapiro

0.67

Activations Density 0.020%

names of specific individuals, potentially related to sports or political figures

proper nouns, particularly names of people

No Comments

No Known Activations