INDEX

Explanations

actions related to physical violence or conflict

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_fs12288-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_12288

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

12,288

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

pton

-0.74

SPONSORED

-0.73

lio

-0.70

ilight

-0.69

Angelo

-0.68

Shin

-0.67

 characterized

-0.63

chart

-0.63

 Introduced

-0.63

 Unloaded

-0.62

POSITIVE LOGITS

 impunity

1.56

 regard

1.34

 regards

1.27

draw

1.21

 utmost

1.17

stood

1.16

 ease

1.16

standing

1.15

 dignity

1.15

 respect

1.10

Activations Density 0.256%

actions related to physical violence or conflict

No Comments

No Known Activations