INDEX

Explanations

phrases that emphasize quantity or a plurality of elements

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

":"/

-0.71

iquette

-0.63

vs

-0.62

%]

-0.54

Leaks

-0.54

oided

-0.52

isse

-0.52

 precedent

-0.52

 marking

-0.51

lash

-0.51

POSITIVE LOGITS

ĪĴ

0.95

 assorted

0.90

etc

0.84

abouts

0.82

£ı

0.79

 contrace

0.77

âĢ¦)

0.77

anything

0.76

milo

0.75

pty

0.71

Activations Density 0.157%

phrases that emphasize quantity or a plurality of elements

No Comments

No Known Activations