INDEX

Explanations

words related to grouping or categorizing objects or concepts

oai_token-act-pair · gpt-3.5-turbo

references to social or organizational groups

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

ocol

-0.68

Nay

-0.66

izzy

-0.62

sten

-0.61

ãĥ¥

-0.60

 everlasting

-0.58

 winters

-0.56

 OLED

-0.56

 Codes

-0.56

 Laws

-0.55

POSITIVE LOGITS

group

3.86

groups

2.70

Group

2.46

 group

2.24

GROUP

2.17

 Groups

1.90

 GROUP

1.88

 Group

1.81

 groups

1.77

 grouping

1.72

Activations Density 0.011%

words related to grouping or categorizing objects or concepts

references to social or organizational groups

No Comments

No Known Activations