INDEX

Explanations

phrases indicating uncertainty or questioning

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

next statement

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

 autorytatywna

-0.66

 nahilalakip

-0.64

 beginnetje

-0.60

AddTagHelper

-0.60

SequentialGroup

-0.58

 propOrder

-0.56

principalTable

-0.55



-0.54

 estekak

-0.54

vacy

-0.54

POSITIVE LOGITS

stateProvider

0.35

XMLSchema

0.34

 viewWillAppear

0.33

encodeWith

0.32

 courtesy

0.32

ついでに

0.30

ViewImports

0.30

 Iglesias

0.30

 from

0.29

sprogramm

0.29

Activations Density 0.069%

phrases indicating uncertainty or questioning

next statement

No Comments

No Known Activations