INDEX

Explanations

attends to the first token related to academic or professional roles from corresponding tokens that indicate personal attributes or identities

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

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.13

1:0.15

2:0.12

3:0.10

4:0.16

5:0.04

6:0.12

7:0.15

Negative Logits

 '\\;'

-0.48

AndEndTag

-0.45

 elettrica

-0.35

 normais

-0.35

хьтан

-0.34

 universitarios

-0.34

 móvel

-0.34

 gynhyrchwyd

-0.33

 صوتيه

-0.33

 giuri

-0.31

POSITIVE LOGITS

 становника

0.39

 AssemblyTitle

0.34

istoitu

0.33

 MonoBehaviour

0.33

헌

0.32

ocirc

0.32

 محفوظة

0.32

 ujednoznacz

0.32

кономі

0.31

vailability

0.30

Activations Density 13.709%

attends to the first token related to academic or professional roles from corresponding tokens that indicate personal attributes or identities

No Comments

No Known Activations

attends to the first token related to academic or professional roles from corresponding tokens that indicate personal attributes or identities

No Comments

No Known Activations