INDEX

Explanations

names of politicians or government officials

oai_token-act-pair · gpt-3.5-turbo

references to specific individuals in positions of authority or governance

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

Brit

-0.73

REF

-0.71

 Siberian

-0.71

lust

-0.69

 Sleeping

-0.67

 sleeper

-0.67

Roaming

-0.67

 Nietzsche

-0.64

 drift

-0.64

AGE

-0.64

POSITIVE LOGITS

 oversees

1.23

 vetoed

1.21

 presided

1.21

 apologized

1.15

 announced

1.15

 convened

1.14

 ordered

1.13

 veto

1.13

 spokesman

1.11

 issued

1.11

Activations Density 0.218%

names of politicians or government officials

references to specific individuals in positions of authority or governance

No Comments

No Known Activations