INDEX

Explanations

attends to tokens containing specific terminologies from mathematical or scientific contexts from standard text 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

0:0.09

1:0.09

2:0.50

3:0.05

4:0.05

5:0.07

6:0.03

7:0.08

Negative Logits

 CommonModule

-0.36

 }}"></

-0.36

 дописавши

-0.36

extAlignment

-0.36

#![

-0.36

 архивлан

-0.34

。】

-0.33

Portail

-0.33

)))),

-0.32

sizeCache

-0.32

POSITIVE LOGITS

apunov

0.45

 EconPapers

0.41

resizingMask

0.41

 actuelles

0.39

 scolaires

0.38

 消息

0.37

 estreno

0.36

Bronnen

0.36

 textStatus

0.36

 alimentaires

0.35

Activations Density 1.017%

attends to tokens containing specific terminologies from mathematical or scientific contexts from standard text tokens

No Comments

No Known Activations