INDEX

Explanations

attends to approval rating-related tokens from corresponding review aggregation 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

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.07

1:0.07

2:0.09

3:0.25

4:0.07

5:0.05

6:0.27

7:0.10

Negative Logits

GGI

-0.34

RenderAtEndOf

-0.33

 Lath

-0.32

 quadrinhos

-0.32

 sintético

-0.31

ături

-0.31

OGND

-0.30

 ſtate

-0.30

afs

-0.30

ósitos

-0.30

POSITIVE LOGITS

ьаж

0.40

TypedDataSet

0.35

istoitu

0.35

DATABASES

0.35

 ISTAT

0.34

ftagPool

0.33

estones

0.32

hidupan

0.32

ChildScrollView

0.31

ніципа

0.31

Activations Density 0.001%

attends to approval rating-related tokens from corresponding review aggregation tokens

No Comments

No Known Activations