INDEX

Explanations

references to individuals or groups involved in actions or events

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

who followed by action

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

ModelState

-0.37

 considerable

-0.35

tvguidetime

-0.34

 nessuna

-0.33

 olyan

-0.32

 several

-0.32

 nějak

-0.32

 tarko

-0.31

 erila

-0.31

 without

-0.31

POSITIVE LOGITS

LookAnd

0.69

 Chwiliwch

0.65

 jScrollPane

0.64

UrlResolution

0.59

acchar

0.59

DockStyle

0.58

PYX

0.58

RegistryLite

0.58

protoc

0.57

StoreMessageInfo

0.57

Activations Density 0.036%

references to individuals or groups involved in actions or events

who followed by action

No Comments

No Known Activations