INDEX

Explanations

names or terms related to individuals or identities

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 @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

regulated

-0.76

soDeliveryDate

-0.71

 resil

-0.71

 compan

-0.68

 lease

-0.65

paren

-0.64

è¦ļéĨĴ

-0.64

 technically

-0.63

sth

-0.63

groups

-0.62

POSITIVE LOGITS

 Patel

1.21

 Hussain

1.18

 Rahman

1.17

 Singh

1.17

 Sharma

1.15

 Chak

1.14

 Ahmed

1.13

 Mohamed

1.12

Sz

1.11

 Krish

1.11

Activations Density 0.262%

names or terms related to individuals or identities

proper nouns, particularly names of people

No Comments

No Known Activations