INDEX

Explanations

phrases related to comparison or emphasis

oai_token-act-pair · gpt-3.5-turbo

connections to magnitude or scale in various contexts

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

orno

-0.78

Cola

-0.76

arter

-0.70

rentices

-0.69

Sem

-0.68

bis

-0.68

ĺ

-0.67

iliate

-0.66

quote

-0.65

arent

-0.65

POSITIVE LOGITS

 magnitude

1.54

 amount

1.41

 enorm

1.40

 extent

1.38

 seriousness

1.37

 lengths

1.36

 severity

1.36

 havoc

1.35

 impact

1.30

 vitri

1.28

Activations Density 0.278%

phrases related to comparison or emphasis

connections to magnitude or scale in various contexts

No Comments

No Known Activations