INDEX

Explanations

terms related to someone's participation or connection in a particular activity or situation

oai_token-act-pair · gpt-3.5-turbo

mentions of involvement 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 @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

flo

-0.66

here

-0.64

 Gleaming

-0.62

ilver

-0.62

leans

-0.61

tted

-0.61

wered

-0.60

und

-0.59

blank

-0.59

adobe

-0.58

POSITIVE LOGITS

 involvement

3.54

 participation

2.15

 complicity

1.80

 affiliation

1.60

involved

1.56

 engagement

1.46

 intervention

1.39

 role

1.36

 presence

1.35

 contribution

1.35

Activations Density 0.017%

terms related to someone's participation or connection in a particular activity or situation

mentions of involvement in various contexts

No Comments

No Known Activations