INDEX

Explanations

attends to tokens that are brand or product names from relevant tokens of the same category

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

0:0.09

1:0.34

2:0.09

3:0.07

4:0.12

5:0.10

6:0.06

7:0.08

Negative Logits

 للاسماء

-0.54

})));

-0.46

 față

-0.45

NUMX

-0.45

 nahilalakip

-0.45

(!__

-0.44

matchCondition

-0.44

}));

-0.43

]));

-0.43

UIControlState

-0.42

POSITIVE LOGITS

ROGEN

0.38

mities

0.33

 nature

0.33

 kone

0.33

らか

0.32

vete

0.32

azine

0.32

got

0.32

洋

0.32

 will

0.31

Activations Density 0.498%

attends to tokens that are brand or product names from relevant tokens of the same category

No Comments

No Known Activations