INDEX

Explanations

politicians and their political affiliations

oai_token-act-pair · gpt-3.5-turbo

opening parentheses in references to political figures and events

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 smugglers

-0.79

 rave

-0.76

 customs

-0.72

 infring

-0.71

 envy

-0.70

 converter

-0.69

 intrinsic

-0.69

 astronomical

-0.68

 superst

-0.67

 Unic

-0.66

POSITIVE LOGITS

Republican

1.27

pictured

1.22

Va

1.17

Fla

1.15

who

1.14

Calif

1.14

TX

1.10

Iowa

1.07

1.06

whose

1.06

Activations Density 0.032%

politicians and their political affiliations

opening parentheses in references to political figures and events

No Comments

No Known Activations