INDEX

Explanations

attends to tokens indicating numeric or identifier values from tokens indicating text or descriptor values

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

0:0.13

1:0.42

2:0.07

3:0.09

4:0.08

5:0.05

6:0.05

7:0.07

Negative Logits

хьтан

-0.41

JsonHelper

-0.35

fjspx

-0.35

XDECREF

-0.33

новниш

-0.33

 dAtA

-0.31

AssignableFrom

-0.30

AlterField

-0.30

 ModelExpression

-0.30

 allAfrica

-0.30

POSITIVE LOGITS

expandindo

0.43

PhysRevD

0.39

amerikanischer

0.39

Environ

0.39

 otomatig

0.35

 تانيه

0.35

 Environ

0.35

 Outro

0.34

stateProvider

0.34

Breite

0.34

Activations Density 0.899%

attends to tokens indicating numeric or identifier values from tokens indicating text or descriptor values

No Comments

No Known Activations