INDEX

Explanations

mentions of political positions or roles, specifically related to the position of "Speaker"

oai_token-act-pair · gpt-3.5-turbo

references to political speakers or positions

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

 Devils

-0.65

urat

-0.65

 Mines

-0.64

arial

-0.63

ster

-0.62

sters

-0.59

iko

-0.59

opia

-0.58

 infinity

-0.58

 Veterinary

-0.57

POSITIVE LOGITS

phone

1.10

 speaker

0.94

person

0.92

hip

0.88

board

0.87

woman

0.85

glass

0.83

 Gingrich

0.79

boards

0.79

loo

0.79

Activations Density 0.037%

mentions of political positions or roles, specifically related to the position of "Speaker"

references to political speakers or positions

No Comments

No Known Activations