INDEX

Explanations

attends to the token "fonts" from math-related tokens

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

0:0.17

1:0.14

2:0.23

3:0.12

4:0.10

5:0.06

6:0.07

7:0.08

Negative Logits

 aDecoder

-0.32

المراجع

-0.29

Przypisy

-0.29

ScopeManager

-0.28

ագրություններ

-0.28

quoise

-0.28

 Defensa

-0.27

erto

-0.27

 yourself

-0.27

ฺ

-0.26

POSITIVE LOGITS

OGND

0.35

 LiveData

0.34

LookAnd

0.33

ьаж

0.31

avyzd

0.29

 MainAxisSize

0.28

'{@

0.28

IntoConstraints

0.28

 Gump

0.28

IBAction

0.28

Activations Density 0.009%

attends to the token "fonts" from math-related tokens

No Comments

No Known Activations