INDEX

Explanations

instances of the word "over" followed by a number indicating intensity or quantity

oai_token-act-pair · gpt-3.5-turbo

phrases related to substantial quantities or frequency of occurrences

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

illary

-0.81

arium

-0.69

 partName

-0.69

UE

-0.67

KE

-0.64

uary

-0.64

Ô

-0.63

UX

-0.61

ero

-0.61

udes

-0.61

POSITIVE LOGITS

tones

0.91

drive

0.88

lord

0.87

haul

0.84

loading

0.82

elin

0.79

population

0.79

hang

0.78

valued

0.77

clock

0.77

Activations Density 0.075%

instances of the word "over" followed by a number indicating intensity or quantity

phrases related to substantial quantities or frequency of occurrences

No Comments

No Known Activations