INDEX

Explanations

references to religion and related concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

Data Type

torch.float32

Hook Point

blocks.8.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

HAEL

-0.86

 Pigs

-0.68

enegger

-0.67

 slate

-0.65

 Wilde

-0.62

BOX

-0.61

 buck

-0.60

schild

-0.60

é¾įå¥ĳå£«

-0.59

BOX

-0.57

POSITIVE LOGITS

igion

1.38

iever

1.32

iance

1.30

iability

1.23

iable

1.19

ativity

1.17

ieving

1.14

atively

1.13

ievers

1.13

atives

1.13

Activations Density 0.007%

references to religion and related concepts

No Comments

No Known Activations