INDEX

Explanations

connection words or conjunctions in text

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

conjunction and followed by connectors

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

 AssemblyVersion

-0.88

Personendaten

-0.81

 صوتيه

-0.76

findpost

-0.75

IntoConstraints

-0.73

 مرئيه

-0.71

 AssemblyTitle

-0.70

 Himo

-0.68

 invokingState

-0.66

 للمعارف

-0.66

POSITIVE LOGITS

<bos>

0.71

“

0.42

by

0.36

zeiti

0.36

AND

0.36

 Verwand

0.32

with

0.31

and

0.31

withstanding

0.31

di

0.31

Activations Density 0.068%

connection words or conjunctions in text

conjunction and followed by connectors

No Comments

No Known Activations