INDEX

Explanations

attends to the word "left" from tokens that reference specific locations or settings

oai_attention-head · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 0-gemmascope-att-16k

Configuration

google/gemma-scope-2b-pt-att/layer_0/width_16k/average_l0_104

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.0.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

0:0.02

1:0.01

2:0.01

3:0.02

4:0.77

5:0.02

6:0.06

7:0.03

Negative Logits

textrm

-0.30

yks

-0.29

 secco

-0.28

 mathvariant

-0.28

اط

-0.28

 hadapan

-0.27

ற்

-0.27

-0.26

 beso

-0.26

ArrayAdapter

-0.26

POSITIVE LOGITS

AddTagHelper

0.57

 مشين

0.56

IndentedString

0.54

 дописавши

0.52

 InputDecoration

0.49

StoryboardSegue

0.48

 gyhoeddwyd

0.47

MigrationBuilder

0.45

Билгалдахарш

0.45

RenderAtEndOf

0.45

Activations Density 0.212%

attends to the word "left" from tokens that reference specific locations or settings

No Comments

No Known Activations