INDEX

Explanations

attends to specific named entities from their corresponding categories or descriptions

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

0:0.20

1:0.48

2:0.06

3:0.04

4:0.05

5:0.03

6:0.04

7:0.07

Negative Logits

jon

-0.33

-0.30

ver

-0.29

…

-0.29

CodedInputStream

-0.28

Personensuche

-0.28

ViewFeatures

-0.28

 समीक्षक

-0.28

...

-0.28

 TextAppearance

-0.28

POSITIVE LOGITS

 itſelf

0.44

 myſelf

0.42

 inoxydable

0.39

 Majefty

0.39

 zubehör

0.37

]='\

0.37

Efq

0.36

erweise

0.36

adecimal

0.36

skosten

0.35

Activations Density 1.061%

attends to specific named entities from their corresponding categories or descriptions

No Comments

No Known Activations