INDEX

Explanations

punctuation and connecting words that indicate relationships or continuations in sentences

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

joining clauses and lists

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

sarung

-0.50

 CreateTagHelper

-0.49

 juries

-0.48

üyor

-0.46

 ainfi

-0.45

 bridesmaid

-0.44

 StatefulWidget

-0.43

 crunchy

-0.43

 jadx

-0.43

erintah

-0.43

POSITIVE LOGITS

 removal

0.56

 identification

0.53

 participation

0.52

 subsequent

0.52

 resulting

0.51

 inability

0.51

 involving

0.50

 consequent

0.50

 issuance

0.48

 incorporation

0.48

Activations Density 0.122%

punctuation and connecting words that indicate relationships or continuations in sentences

joining clauses and lists

No Comments

No Known Activations