INDEX

Explanations

attends to tokens marked with double asterisks from tokens marked with double square brackets that appear later in the sequence

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

1:0.06

2:0.05

3:0.07

4:0.07

5:0.04

6:0.54

7:0.07

Negative Logits

 createSlice

-0.35

rungsseite

-0.34

 surla

-0.34

 proprement

-0.33

ScopeManager

-0.33

UnitTesting

-0.33

xhr

-0.32

 kasarigan

-0.32

Tracce

-0.32

BlockPos

-0.32

POSITIVE LOGITS

 sebag

0.28

ándola

0.26

 aang

0.26

ouwen

0.26

 maat

0.26

hages

0.26

ándolos

0.26

 caramelo

0.25

 YouTube

0.25

ilitar

0.25

Activations Density 0.074%

attends to tokens marked with double asterisks from tokens marked with double square brackets that appear later in the sequence

No Comments

No Known Activations