INDEX

Explanations

terms related to fire, burning, or risks associated with combustion

oai_token-act-pair · gpt-4o-mini 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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

PLAY

-0.78

ASED

-0.68

information

-0.64

MAR

-0.64

Copyright

-0.63

 unequ

-0.62

Ô

-0.61

merga

-0.60

 allowances

-0.59

Xi

-0.59

POSITIVE LOGITS

shire

1.11

1.05

sburg

1.04

icz

1.03

ian

1.03

0.99

ski

0.97

sie

0.97

sky

0.95

ersen

0.94

Activations Density 0.144%

terms related to fire, burning, or risks associated with combustion

No Comments

No Known Activations