INDEX

Explanations

mentions of a specific news network

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

rians

-0.82

akeru

-0.78

 inval

-0.75

rian

-0.74

ictions

-0.71

acterial

-0.70

rogens

-0.70

ansom

-0.69

adding

-0.69

assy

-0.68

POSITIVE LOGITS

conn

1.50

 News

0.98

News

0.94

hawk

0.89

hound

0.89

woods

0.86

croft

0.85

fox

0.83

cat

0.83

FOX

0.83

Activations Density 0.519%

mentions of a specific news network

No Comments

No Known Activations