INDEX

Explanations

attends to tokens featuring "radi," "pseud," and "conspir" from tokens 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 @ 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

0:0.05

1:0.01

2:0.02

3:0.01

4:0.63

5:0.19

6:0.01

7:0.04

Negative Logits

 springfox

-0.35

 ModelExpression

-0.33

 NUKAT

-0.32

 whole

-0.30

EndContext

-0.28

 ostavi

-0.28

 nearest

-0.28

IntoConstraints

-0.27

psack

-0.27

 استنادى

-0.27

POSITIVE LOGITS

 LoginComponent

0.33

XmlRootElement

0.33

 PyTuple

0.31

MLLoader

0.31

ParallelGroup

0.30

genza

0.29

섹

0.29

wahati

0.29

fillText

0.29

eload

0.29

Activations Density 2.505%

attends to tokens featuring "radi," "pseud," and "conspir" from tokens that appear later in the sequence

No Comments

No Known Activations