INDEX

Explanations

phrases indicating value or importance

oai_token-act-pair · gpt-3.5-turbo

expressions related to value or worth

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

alde

-0.61

ablishment

-0.61

ce

-0.58

ads

-0.58

estic

-0.57

dog

-0.57

di

-0.56

eret

-0.56

REAM

-0.55

 Pressure

-0.54

POSITIVE LOGITS

 worth

4.00

 Worth

2.31

worth

2.01

 worthwhile

1.97

 valued

1.59

 worthy

1.53

 costing

1.26

worthy

1.25

 priceless

1.20

 deserving

1.16

Activations Density 0.015%

phrases indicating value or importance

expressions related to value or worth

No Comments

No Known Activations