INDEX

Explanations

names of individuals

oai_token-act-pair · gpt-3.5-turbo

proper nouns, particularly names and initials

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

å§«

-0.97

ãĥ¼ãĤ¯

-0.77

urat

-0.76

 Catalonia

-0.70

ModLoader

-0.69

 âĸº

-0.68

âĶĢâĶĢ

-0.68

EVA

-0.67

 Pradesh

-0.66

 Lanka

-0.66

POSITIVE LOGITS

aylor

0.79

iggs

0.76

isner

0.75

antz

0.73

verson

0.71

elson

0.69

aney

0.68

arkin

0.68

ahl

0.68

horn

0.67

Activations Density 0.222%

names of individuals

proper nouns, particularly names and initials

No Comments

No Known Activations