INDEX

Explanations

statements indicating the existence or condition of entities

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

ado

-0.71

oids

-0.69

ties

-0.68

udeau

-0.67

ths

-0.67

 tendency

-0.67

 rant

-0.66

 penchant

-0.64

olor

-0.64

urous

-0.63

POSITIVE LOGITS

 properly

1.03

 able

1.01

 adequately

0.93

 safe

0.92

abella

0.88

Ħ¢

0.88

 aware

0.88

 capable

0.88

 ready

0.88

 correct

0.85

Activations Density 0.110%

statements indicating the existence or condition of entities

No Comments

No Known Activations