INDEX

Explanations

phrases related to specific organizational names and locations

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

wcsstore

-0.74

 Luthor

-0.70

ths

-0.67

chrom

-0.65

IGHTS

-0.63

à¥

-0.61

ãĥīãĥ©

-0.60

ods

-0.60

à©

-0.59

BuyableInstoreAndOnline

-0.59

POSITIVE LOGITS

pieces

0.88

 pamph

0.83

 fielder

0.78

ologies

0.77

tyard

0.76

stadt

0.76

eous

0.72

jer

0.72

piece

0.72

 staffed

0.70

Activations Density 0.034%

phrases related to specific organizational names and locations

No Comments

No Known Activations