INDEX

Explanations

phrases indicating agreement or alignment with a position or statement

oai_token-act-pair · gpt-3.5-turbo

expressions of agreement

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

oufl

-0.74

 glands

-0.73

crow

-0.71

 Tycoon

-0.69

 typh

-0.69

vas

-0.66

 Blooming

-0.65

 tremend

-0.60

brid

-0.58

 Mongolia

-0.58

POSITIVE LOGITS

rences

0.84

 unanimously

0.82

reements

0.80

ipeg

0.80

agree

0.78

ably

0.76

lihood

0.76

reement

0.75

 agreement

0.74

ettle

0.73

Activations Density 0.024%

phrases indicating agreement or alignment with a position or statement

expressions of agreement

No Comments

No Known Activations