INDEX

Explanations

phrases related to official statements or reports from representatives or sources

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

statement by proper noun

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

featureID

-0.64

uxxxx

-0.57



-0.55

 faſt

-0.54

 OnTrigger

-0.53

InputTagHelper

-0.52

ngrx

-0.50

 betweenstory

-0.50

AddTagHelper

-0.49

 Chwiliwch

-0.48

POSITIVE LOGITS

 initComponents

0.57

the

0.49

 official

0.48

official

0.43

 AssemblyCompany

0.42

 united

0.39

 unnamed

0.36

Official

0.36

mappedBy

0.35

 ufficiale

0.35

Activations Density 0.021%

phrases related to official statements or reports from representatives or sources

statement by proper noun

No Comments

No Known Activations