INDEX

Explanations

phrases related to procedures and steps, such as assessing, evaluating, and being granted access

oai_token-act-pair · gpt-3.5-turbo

the conjunction "and" used in various contexts, indicating addition or connection

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

 Saying

-0.78

ukemia

-0.73

maxwell

-0.70

 Drawing

-0.70

ibble

-0.70

onica

-0.69

inch

-0.67

reetings

-0.66

warm

-0.63

mma

-0.62

POSITIVE LOGITS

 reused

1.48

 redistributed

1.40

 replaced

1.40

 eaten

1.36

 taxed

1.31

 analyzed

1.31

 inspected

1.31

 analysed

1.30

 marketed

1.29

 thrown

1.29

Activations Density 0.207%

phrases related to procedures and steps, such as assessing, evaluating, and being granted access

the conjunction "and" used in various contexts, indicating addition or connection

No Known Activations