INDEX

Explanations

attends to various tokens marked as important or significant from contextual elements discussed 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

0:0.19

1:0.17

2:0.10

3:0.12

4:0.10

5:0.05

6:0.07

7:0.16

Negative Logits

 gesamte

-0.26

GoogleApiClient

-0.26

postDelayed

-0.26

ότε

-0.25

Next

-0.25

 Next

-0.25

vernight

-0.25

bov

-0.24

grouped

-0.23

 Gesamt

-0.23

POSITIVE LOGITS

GEBURTSDATUM

0.46

featureID

0.45

Sucesor

0.43

principalTable

0.42

UserScript

0.41

lapsingToolbar

0.41

 bezeichneter

0.40

 متعلقه

0.40

 ModelExpression

0.40

parsedMessage

0.39

Activations Density 0.297%

attends to various tokens marked as important or significant from contextual elements discussed later in the sequence

No Comments

No Known Activations