INDEX

Explanations

words related to leadership and authority in governmental or organizational contexts

oai_token-act-pair · gpt-3.5-turbo

terms associated with entities or participants in a context, particularly focusing on nouns

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

 numbered

-0.66

 photos

-0.65

 imagery

-0.64

 estimate

-0.62

 Sonic

-0.62

 streams

-0.62

 bins

-0.61

 presumably

-0.60

 photographs

-0.60

 libraries

-0.60

POSITIVE LOGITS

ente

4.94

ent

1.47

entin

1.34

ento

1.29

enting

1.21

enza

1.17

encia

1.17

ENT

1.16

enta

1.16

ante

1.16

Activations Density 0.021%

words related to leadership and authority in governmental or organizational contexts

terms associated with entities or participants in a context, particularly focusing on nouns

No Comments

No Known Activations