INDEX

Explanations

names or entities related to individuals, especially when combined with actions or locations

oai_token-act-pair · gpt-3.5-turbo

proper names, particularly of individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

ctors

-0.73

aneously

-0.67

 Presidents

-0.66

acts

-0.65

cers

-0.63

 presidents

-0.63

fare

-0.63

ously

-0.63

istically

-0.62

condition

-0.62

POSITIVE LOGITS

mith

1.56

hift

1.42

peed

1.34

chool

1.32

hip

1.32

hirt

1.32

heet

1.30

pring

1.30

erver

1.30

creen

1.29

Activations Density 0.186%

names or entities related to individuals, especially when combined with actions or locations

proper names, particularly of individuals

No Comments

No Known Activations