INDEX

Explanations

attends to the token "td" from corresponding "height" attributes in earlier tokens that are related to table structures

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

0:0.34

1:0.18

2:0.07

3:0.14

4:0.06

5:0.06

6:0.05

7:0.06

Negative Logits

CodedInputStream

-0.35

хьтан

-0.30

 Monfieur

-0.29

 Roskov

-0.29

posedge

-0.29

Efq

-0.29

 jeter

-0.28

werd

-0.28

oa̍t

-0.27

 jette

-0.27

POSITIVE LOGITS

crollView

0.27

WriteTagHelper

0.26

uxxxx

0.25

 kanssa

0.25

 kasarigan

0.25

 оригіналу

0.24

medriver

0.24

graduates

0.24

addafi

0.24

ogues

0.23

Activations Density 0.075%

attends to the token "td" from corresponding "height" attributes in earlier tokens that are related to table structures

No Comments

No Known Activations