INDEX

Explanations

attends to the conclusion-related tokens marked as "that" from the preceding tokens marked as "concluded."

oai_attention-head · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 0-gemmascope-att-16k

Configuration

google/gemma-scope-2b-pt-att/layer_0/width_16k/average_l0_104

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.0.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

0:0.44

1:0.02

2:0.02

3:0.02

4:0.35

5:0.04

6:0.02

7:0.06

Negative Logits

 itſelf

-0.53

 whoſe

-0.46

PreferredItem

-0.45

 ſhe

-0.45

ArrowToggle

-0.45

 ſhould

-0.45

 becauſe

-0.44

例文帳に追加

-0.43

Efq

-0.43

ſelves

-0.43

POSITIVE LOGITS

сла

0.33

bé

0.32

arti

0.31

0.30

neh

0.30

Mere

0.30

 telefon

0.30

>(_

0.30

orod

0.29

ära

0.29

Activations Density 0.064%

attends to the conclusion-related tokens marked as "that" from the preceding tokens marked as "concluded."

No Comments

No Known Activations