INDEX

Explanations

attends to pairs of tokens describing substantial historical or political concepts from tokens implying factual statements or projections

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

0:0.07

1:0.08

2:0.07

3:0.13

4:0.12

5:0.02

6:0.29

7:0.19

Negative Logits

enderror

-0.31

ayah

-0.31

 NSCoder

-0.30

 חיצוניים

-0.30

 ocul

-0.29

 originali

-0.29

sar

-0.28

paž

-0.28

 quæ

-0.28

lectable

-0.28

POSITIVE LOGITS

脚注の使い方

0.39

 allegedly

0.36

CloseOperation

0.34

XmlAccessorType

0.34

WebVitals

0.32

Referencer

0.32

SequentialGroup

0.31

 supuestamente

0.31

 SwiftUI

0.30

RenderAtEndOf

0.30

Activations Density 0.327%

attends to pairs of tokens describing substantial historical or political concepts from tokens implying factual statements or projections

No Comments

No Known Activations