INDEX

Explanations

attends to the continuation of actions or states from a variety of contexts marked by "to."

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

0:0.78

1:0.01

2:0.01

3:0.01

4:0.12

5:0.02

6:0.01

7:0.02

Negative Logits

Vidite

-0.51

TagMode

-0.48

EDEFAULT

-0.46

IndentedString

-0.45

 ModelExpression

-0.44

---*/

-0.43

ItemBackground

-0.42



-0.40

 Numerade

-0.39

Rujuakan

-0.39

POSITIVE LOGITS

ори

0.28

 reconstruct

0.28

 vời

0.27

 authenticate

0.27

GOR

0.27

ough

0.27

alth

0.27

 ratify

0.27

 Zufall

0.26

Oss

0.26

Activations Density 0.028%

attends to the continuation of actions or states from a variety of contexts marked by "to."

No Comments

No Known Activations