INDEX

Explanations

attends to tokens related to validity or conditions from tokens discussing characteristics or existing states

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

0:0.05

1:0.10

2:0.09

3:0.12

4:0.40

5:0.05

6:0.07

7:0.08

Negative Logits

 Италијани

-0.29

rungsseite

-0.28

Vidite

-0.26

uxxxx

-0.25

 beurre

-0.25

 Tubes

-0.25

 nahilalakip

-0.24

postsleuth

-0.24

ibus

-0.24

 Settle

-0.24

POSITIVE LOGITS

WriteTagHelper

0.35

原始内容存档于

0.27

 发表于

0.27

hdashline

0.26

inetics

0.24

 anymore

0.24

 originalen

0.24

ContentAsync

0.23

NCBI

0.23

 […]

0.23

Activations Density 0.817%

attends to tokens related to validity or conditions from tokens discussing characteristics or existing states

No Comments

No Known Activations