INDEX

Explanations

attends to code termination markers from code beginning tokens

oai_attention-head · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 0-gemmascope-att-16k

Configuration

google/gemma-scope-2b-pt-att/layer_0/width_16k/average_l0_104

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.0.attn.hook_z

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.02

1:0.65

2:0.16

3:0.02

4:0.03

5:0.04

6:0.03

7:0.02

Negative Logits

end

-0.44

end

-0.41

-0.36

End

-0.33

“

-0.33

-0.32

in

-0.31

ex

-0.30

POSITIVE LOGITS

 Shaksp

0.54

Efq

0.52

 photolibrary

0.52

 Shakspeare

0.52

 Whigs

0.51

 Mahomet

0.51

+#+#

0.50

 pleaſure

0.50

 Monfieur

0.49

 Jefus

0.49

Activations Density 0.026%

attends to code termination markers from code beginning tokens

No Comments

No Known Activations