INDEX

Explanations

names of individuals

oai_token-act-pair · gpt-3.5-turbo

proper nouns and names of people or entities in the text

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

mean

-0.68

paren

-0.66

 underestimate

-0.63

Talking

-0.62

 Mean

-0.62

 Yesterday

-0.62

 Speaking

-0.62

advertisement

-0.61

 Talking

-0.60

hap

-0.60

POSITIVE LOGITS

 resumed

1.60

 reverted

1.47

 retreated

1.37

 returned

1.36

 withdrew

1.36

 regained

1.30

 proceeded

1.30

 reopened

1.27

 ceased

1.20

 disappeared

1.19

Activations Density 0.267%

names of individuals

proper nouns and names of people or entities in the text

No Comments

No Known Activations