INDEX

Explanations

punctuated sentences that convey significant or conclusive information

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

one can

np_acts-logits-general · gemini-2.5-flash-lite

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Numerade

-0.57

featureID

-0.55

setVerticalGroup

-0.51

tvguidetime

-0.51

tagHelperRunner

-0.50

 للاسماء

-0.49

addPreferredGap

-0.48

ReusableCell

-0.47

twimg

-0.47

WriteAttribute

-0.47

POSITIVE LOGITS

 ujednoznacz

0.44

曖昧さ回避

0.39

htä

0.39

 Ведь

0.39

+#+

0.37

Lời

0.35

 виправивши

0.35

ẨM

0.34

 Play

0.34

But

0.33

Activations Density 0.107%

punctuated sentences that convey significant or conclusive information

one can

No Comments

No Known Activations

punctuated sentences that convey significant or conclusive information

one can

No Comments

No Known Activations