INDEX

Explanations

phrases associated with numerical data and specifics surrounding various topics

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

bucks

-0.75

ãĥŁ

-0.74

é¾

-0.73

video

-0.72

Ô

-0.69

body

-0.68

potion

-0.68

chens

-0.66

ãĥīãĥ©

-0.65

factor

-0.65

POSITIVE LOGITS

 both

1.13

 various

1.00

 numerous

0.86

 several

0.85

 each

0.78

 varying

0.76

 embassies

0.75

both

0.75

 innumerable

0.74

 Various

0.74

Activations Density 0.263%

phrases associated with numerical data and specifics surrounding various topics

No Comments

No Known Activations