INDEX

Explanations

attends to grade-related tokens from paired grade-related tokens that come after

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

0:0.13

1:0.24

2:0.13

3:0.06

4:0.14

5:0.09

6:0.05

7:0.12

Negative Logits

parsedMessage

-0.39

 estimés

-0.38

findpost

-0.37

 ChromeDriver

-0.37

#+#

-0.37

VersionUID

-0.37

 createContext

-0.35

 насељу

-0.35

 GenerationType

-0.35

\{\\

-0.35

POSITIVE LOGITS

’.”

0.20

.”

0.20

viamo

0.19

 valley

0.19

standing

0.19

ણી

0.19

.’”

0.19

tap

0.19

 fleuve

0.19

ilie

0.18

Activations Density 0.058%

attends to grade-related tokens from paired grade-related tokens that come after

No Comments

No Known Activations