INDEX

Explanations

attends to specific action words from phrases that involve user interaction or method calls

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

0:0.10

1:0.12

2:0.10

3:0.15

4:0.16

5:0.03

6:0.17

7:0.12

Negative Logits

IContainer

-0.33

 judiciaire

-0.28

WriteTagHelper

-0.28

خوری

-0.28

tan

-0.26

談社

-0.25

HCR

-0.25

tan

-0.25

HasAnnotation

-0.25

Kar

-0.24

POSITIVE LOGITS

Viitteet

0.38

\{\\

0.33

Enllaces

0.31

脚注の使い方

0.30

 Audiodateien

0.30

enumi

0.29

 linkovi

0.28

:^{

0.28

eien

0.28

esimo

0.28

Activations Density 0.001%

attends to specific action words from phrases that involve user interaction or method calls

No Comments

No Known Activations