INDEX

Explanations

titles and positions related to advanced academic or professional roles

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

NetMessage

-0.73

 birthplace

-0.68

 cafes

-0.66

 extremes

-0.62

\/\/

-0.62

erella

-0.60

 whim

-0.58

Peb

-0.58

pha

-0.57

 Pengu

-0.57

POSITIVE LOGITS

zinski

0.81

endment

0.75

orer

0.69

cture

0.68

awei

0.67

hips

0.66

ariat

0.66

 Strategic

0.65

Ake

0.64

riel

0.64

Activations Density 0.072%

titles and positions related to advanced academic or professional roles

No Comments

No Known Activations