INDEX

Explanations

phrases related to reports and statements made in various contexts

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

imilar

-0.69

ogenic

-0.67

ãĥİ

-0.66

estern

-0.64

 sickness

-0.63

icio

-0.61

 Birthday

-0.61

ãĥĩ

-0.56

animate

-0.56

 stump

-0.55

POSITIVE LOGITS

 quoting

0.81

 adding

0.74

 bluntly

0.72

 citing

0.70

lege

0.65

0.64

 omin

0.64

 cris

0.63

 referring

0.63

 noting

0.61

Activations Density 0.091%

phrases related to reports and statements made in various contexts

No Comments

No Known Activations