INDEX

Explanations

attends to the tokens "trip," "hike," and "walk" from tokens that imply activities described earlier

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

0:0.09

1:0.08

2:0.09

3:0.42

4:0.08

5:0.03

6:0.07

7:0.09

Negative Logits

irmer

-0.28

 jesu

-0.25

jerne

-0.24

 Decke

-0.24

 paš

-0.24

 исправ

-0.23

med

-0.23

tori

-0.23

en

-0.23

 comod

-0.23

POSITIVE LOGITS

 propOrder

0.47

jsxFileName

0.47

 doInBackground

0.43

]--;

0.43

 Paglinawan

0.41

NOPQRST

0.41

tvguidetime

0.41

EndInit

0.40

WaitGroup

0.40

reportWebVitals

0.39

Activations Density 0.380%

attends to the tokens "trip," "hike," and "walk" from tokens that imply activities described earlier

No Comments

No Known Activations