INDEX

Explanations

prepositions and phrases that indicate location or participation in actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

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

 ProtoMessage

-0.55

 informée

-0.52

 Lightboxes

-0.52

Становништво

-0.52

Demografie

-0.49

 uſed

-0.48

 nakalista

-0.48

ſchen

-0.47

нгред

-0.47

 wireType

-0.47

POSITIVE LOGITS

 mergeFrom

0.43

brave

0.35

dou

0.34

らかに

0.34

 Brave

0.34

 elegantly

0.33

 expertly

0.33

mă

0.32

 söz

0.32

Alike

0.31

Activations Density 0.026%

prepositions and phrases that indicate location or participation in actions

No Comments

No Known Activations