INDEX

Explanations

details related to statistics and geographical descriptions

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Ragnar

-0.61

oki

-0.60

 contrasted

-0.58

 Hutchinson

-0.57

 Guilty

-0.57

NEO

-0.56

 contagious

-0.56

Downloadha

-0.55

 reunited

-0.55

 Sturgeon

-0.55

POSITIVE LOGITS

 ones

1.26

 counterparts

0.83

 thereof

0.83

 versa

0.82

 counterpart

0.80

itud

0.73

 Ones

0.72

rants

0.67

phabet

0.66

quarters

0.65

Activations Density 0.500%

details related to statistics and geographical descriptions

No Comments

No Known Activations