INDEX

Explanations

keywords related to the purpose or intention behind actions or objects

oai_token-act-pair · gpt-3.5-turbo

phrases related to purpose and intention

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

culosis

-0.71

thumbnails

-0.69

 weeds

-0.67

etts

-0.65

 Thumbnails

-0.63

 Leopard

-0.62

 juven

-0.62

hani

-0.62

 Polo

-0.61

UGE

-0.60

POSITIVE LOGITS

ful

1.33

fulness

0.97

fully

0.95

lessness

0.95

tenance

0.94

ality

0.93

lessly

0.92

teness

0.85

FUL

0.84

fare

0.82

Activations Density 0.029%

keywords related to the purpose or intention behind actions or objects

phrases related to purpose and intention

No Comments

No Known Activations