INDEX

Explanations

attends to tokens related to "equals" operations from tokens indicating a type check or view

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

0:0.01

1:0.02

2:0.07

3:0.64

4:0.02

5:0.04

6:0.09

7:0.07

Negative Logits

 unknownFields

-0.45

zuführen

-0.37

errHandler

-0.36

ÁG

-0.36

SSON

-0.36

assumption

-0.36

ratulations

-0.35

ÑA

-0.35

IFORNIA

-0.35

her

-0.35

POSITIVE LOGITS

IntoConstraints

0.38

 Whigs

0.37

 Oedipus

0.36

 myſelf

0.36

 gneiss

0.35

 ſche

0.35

 Infór

0.35

 Etrus

0.34

 uſed

0.34

 ſtate

0.34

Activations Density 0.188%

attends to tokens related to "equals" operations from tokens indicating a type check or view

No Comments

No Known Activations