INDEX

Explanations

terms associated with making decisions or taking action

oai_token-act-pair · gpt-3.5-turbo

instances of the article "a" and its variations 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 @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

anwhile

-0.78

 horizont

-0.74

inx

-0.63

ifles

-0.61

 distances

-0.56

aten

-0.56

arna

-0.56

 seekers

-0.55

makers

-0.55

oran

-0.54

POSITIVE LOGITS

sense

0.84

 impression

0.74

 splash

0.74

 distinction

0.74

 dent

0.72

 contribution

0.68

 difference

0.67

 sense

0.66

Wan

0.65

ument

0.65

Activations Density 0.169%

terms associated with making decisions or taking action

instances of the article "a" and its variations in various contexts

No Comments

No Known Activations