INDEX

Explanations

attends to general negative or neutral statements from positive or self-referential expressions

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

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.24

1:0.21

2:0.17

3:0.09

4:0.06

5:0.02

6:0.07

7:0.11

Negative Logits

--

-0.27

-0.24

 kaynağından

-0.24

-0.23

../../../

-0.23

ीन

-0.23

 Commission

-0.23

timm

-0.23

mousedown

-0.22

POSITIVE LOGITS

 beginnetje

0.44

DockStyle

0.43

 itſelf

0.42

Personendaten

0.42

findpost

0.41

 <=",

0.39

richTextPanel

0.37

ſelf

0.36

èdia

0.35

Efq

0.34

Activations Density 0.189%

attends to general negative or neutral statements from positive or self-referential expressions

No Comments

No Known Activations