INDEX

Explanations

instances of the word "Clinton"

oai_token-act-pair · gpt-3.5-turbo

references to Hillary Clinton

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 stub

-0.71

»Ĵ

-0.70

 realise

-0.69

dash

-0.67

 fitt

-0.66

LU

-0.66

uni

-0.65

 veter

-0.64

 beard

-0.63

AST

-0.63

POSITIVE LOGITS

 Clinton

3.51

Clinton

3.18

 Clintons

2.65

clinton

2.40

 Hillary

2.39

 Rodham

2.26

Hillary

2.24

INTON

2.15

 Sanders

1.89

 Podesta

1.75

Activations Density 0.031%

instances of the word "Clinton"

references to Hillary Clinton

No Comments

No Known Activations