INDEX

Explanations

names with specific professions or locations

oai_token-act-pair · gpt-3.5-turbo

proper nouns and names within a text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ciplinary

-0.82

orthern

-0.81

icing

-0.79

abase

-0.79

raints

-0.77

raint

-0.73

luent

-0.72

cation

-0.72

historic

-0.71

forcement

-0.70

POSITIVE LOGITS

SON

1.71

MAN

1.57

LEY

1.56

TON

1.56

STE

1.55

EY

1.53

WOOD

1.53

BER

1.51

ELL

1.47

VILLE

1.46

Activations Density 0.121%

names with specific professions or locations

proper nouns and names within a text

No Comments

No Known Activations