INDEX

Explanations

instances of the word "trap" and related variations

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

edly

-0.83

sis

-0.76

issance

-0.75

Merit

-0.70

eda

-0.67

cox

-0.66

lihood

-0.65

played

-0.62

soever

-0.62

 normalized

-0.61

POSITIVE LOGITS

door

1.55

doors

1.34

finding

1.00

idon

0.84

trap

0.84

 traps

0.80

 trap

0.78

 Hole

0.76

ERY

0.72

 Trap

0.72

Activations Density 0.007%

instances of the word "trap" and related variations

No Comments

No Known Activations