INDEX

Explanations

names of individuals involved in news stories or events

oai_token-act-pair · gpt-3.5-turbo

proper nouns, particularly names of individuals mentioned in criminal contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

 mathemat

-0.82

 guaranteeing

-0.81

cffff

-0.81

 optim

-0.77

 forecasting

-0.75

 bookmark

-0.74

 charism

-0.73

pmwiki

-0.72

pse

-0.71

 conclud

-0.70

POSITIVE LOGITS

Doe

1.24

 Tsarnaev

1.07

 Zimmerman

1.00

 Bundy

0.96

 Hernandez

0.92

 Ramirez

0.92

 Paddock

0.91

 Martinez

0.87

 Rodriguez

0.87

 Nguyen

0.86

Activations Density 0.428%

names of individuals involved in news stories or events

proper nouns, particularly names of individuals mentioned in criminal contexts

No Comments

No Known Activations