INDEX

Explanations

phrases indicating the utilization of tools or methods

oai_token-act-pair · gpt-3.5-turbo

instances of the word "using" and similar forms, indicating processes or methods employed in various contexts

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

aughs

-0.84

orne

-0.75

tenance

-0.68

rament

-0.65

vals

-0.65

gow

-0.63

cend

-0.61

ologne

-0.60

rick

-0.59

amaz

-0.57

POSITIVE LOGITS

 using

3.12

 utilizing

2.41

 Using

2.15

Using

1.99

using

1.99

 employing

1.92

 relying

1.74

 leveraging

1.63

 incorporating

1.56

 exploiting

1.55

Activations Density 0.040%

phrases indicating the utilization of tools or methods

instances of the word "using" and similar forms, indicating processes or methods employed in various contexts

No Comments

No Known Activations