INDEX

Explanations

the name "Merkel"

oai_token-act-pair · gpt-3.5-turbo

mentions of Angela Merkel

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

aping

-0.76

GROUND

-0.73

orage

-0.71

avery

-0.70

aped

-0.69

aways

-0.69

Methods

-0.69

PASS

-0.68

RECT

-0.68

apes

-0.67

POSITIVE LOGITS

 Merkel

1.31

 bund

0.89

 Nieto

0.82

 chancellor

0.80

baum

0.79

 Tanz

0.76

enstein

0.75

olini

0.72

mann

0.72

 Jinping

0.72

Activations Density 0.005%

the name "Merkel"

mentions of Angela Merkel

No Comments

No Known Activations