INDEX

Explanations

references to work or employment

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

 Ukrain

-0.99

olic

-0.85

 constitu

-0.69

emonic

-0.64

ique

-0.63

pha

-0.62

Sov

-0.62

 Mold

-0.62

 QUEST

-0.61

laus

-0.60

POSITIVE LOGITS

aday

1.25

station

1.11

day

1.08

days

1.06

 ethic

0.99

week

0.97

hour

0.96

night

0.95

bench

0.94

lifting

0.93

Activations Density 0.033%

references to work or employment

No Comments

No Known Activations