INDEX

Explanations

attends to text-related tokens from respective UI element tokens

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 10-gemmascope-att-16k

Configuration

google/gemma-scope-2b-pt-att/layer_10/width_16k/average_l0_70

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.10.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

0:0.06

1:0.20

2:0.07

3:0.05

4:0.23

5:0.23

6:0.05

7:0.06

Negative Logits

AndEndTag

-0.61

 resourceCulture

-0.43

RetentionPolicy

-0.41

 فريبيس

-0.40

RunAsync

-0.39

ExecuteReader

-0.38

 snippetHide

-0.38

(!__

-0.37

 للاسماء

-0.37

POSITIVE LOGITS

 règne

0.23

 assolu

0.22

Rè

0.22

 önem

0.22

 times

0.21

 intéress

0.21

faktor

0.21

land

0.21

 forma

0.21

GGLE

0.20

Activations Density 0.043%

attends to text-related tokens from respective UI element tokens

No Comments

No Known Activations