INDEX

Explanations

phrases emphasizing miscellaneous or vague 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

geon

-0.72

 largeDownload

-0.69

oir

-0.67

iary

-0.66

-0.65

ramid

-0.65

oux

-0.64

ICO

-0.62

onz

-0.62

geons

-0.62

POSITIVE LOGITS

 happening

1.25

 happen

1.22

 happened

1.12

 happens

0.97

 happ

0.92

 transpired

0.85

 Happ

0.84

folk

0.82

 pertaining

0.81

 occurring

0.78

Activations Density 0.040%

phrases emphasizing miscellaneous or vague elements

No Comments

No Known Activations