INDEX

Explanations

words related to enhancing or adding to something

oai_token-act-pair · gpt-3.5-turbo

terms related to complementary or supplementary relationships

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Records

-0.69

urg

-0.67

 Pictures

-0.65

 Murd

-0.64

 unsafe

-0.64

 Despair

-0.62

 Rats

-0.61

violent

-0.61

lees

-0.61

yo

-0.60

POSITIVE LOGITS

 complement

3.97

 compliment

2.31

 complementary

2.27

plement

1.70

 comple

1.57

 augment

1.37

 compliments

1.37

 complimentary

1.31

 synerg

1.28

 supplement

1.24

Activations Density 0.029%

words related to enhancing or adding to something

terms related to complementary or supplementary relationships

No Comments

No Known Activations