INDEX

Explanations

instances of the word "with" indicating associations or relationships

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 6-gemmascope-mlp-16k

Configuration

google/gemma-scope-2b-pt-mlp/layer_6/width_16k/average_l0_133

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.6.hook_mlp_out

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

Negative Logits

 isother

-0.56

momix

-0.56

CTYPE

-0.55

CHECKED

-0.54

enk

-0.52

říklad

-0.52

Matte

-0.52

TAB

-0.52

tab

-0.51

 alongside

-0.50

POSITIVE LOGITS

 mergeFrom

0.70

 propOrder

0.63

 Redacción

0.62

 ويكيميديا

0.61

oa̍t

0.61

principalColumn

0.60

 noDo

0.59

 Wicidata

0.59

BeginContext

0.59

 &___

0.58

Activations Density 0.131%

instances of the word "with" indicating associations or relationships

No Comments

No Known Activations