INDEX

Explanations

locations or regions referenced in a document

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

wagon

-0.59

itle

-0.55

 surpr

-0.55

%%

-0.55

awa

-0.53

 evolves

-0.53

ALSE

-0.53

 summed

-0.52

ens

-0.51

 rightfully

-0.51

POSITIVE LOGITS

 lieu

1.27

 accordance

1.11

 spite

1.04

 conjunction

1.03

 order

1.03

patient

0.97

 favor

0.95

clus

0.93

 vitro

0.93

humane

0.93

Activations Density 0.400%

locations or regions referenced in a document

No Comments

No Known Activations