INDEX

Explanations

attends to instances of "from" related to return or coming back from diverse scenarios, from related contextual tokens

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

0:0.12

1:0.14

2:0.16

3:0.10

4:0.05

5:0.02

6:0.05

7:0.32

Negative Logits

 "");

-0.28

 '');

-0.27

 CURLOPT

-0.25

 />";

-0.25

Tikang

-0.24

'',

-0.24

\"");

-0.24

ConnectionState

-0.24

 nakalista

-0.23

TestClass

-0.23

POSITIVE LOGITS

parsedMessage

0.30

 perm

0.23

tadiene

0.23

auré

0.22

dification

0.22

 Autorisations

0.21

 propOrder

0.21

WireFormatLite

0.21

FormTagHelper

0.21

 ruolo

0.21

Activations Density 0.127%

attends to instances of "from" related to return or coming back from diverse scenarios, from related contextual tokens

No Comments

No Known Activations