INDEX

Explanations

instances of high levels of activity or turn-taking in conversation

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

code or programming terms

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

 ört

-0.39

MockBean

-0.38

StartTag

-0.38

COLN

-0.36

OrNil

-0.36

bcryptjs

-0.35

XMLSchema

-0.34

 armées

-0.34

tagez

-0.34

 çap

-0.33

POSITIVE LOGITS

 StatefulWidget

0.58

tagHelperRunner

0.57

principalTable

0.56

.*")]

0.45

 StatelessWidget

0.45

fromnode

0.45

Parcelize

0.43

}-*/;

0.43

 snippetHide

0.42

 interloc

0.42

Activations Density 0.087%

instances of high levels of activity or turn-taking in conversation

code or programming terms

No Comments

No Known Activations