INDEX

Explanations

requests for information or feedback from users

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

users, fans, recognize

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

UseVisualStyle

-0.37

AxisAlignment

-0.33

 <=",

-0.32

 INPUTS

-0.31

Schedulers

-0.31

PostExecute

-0.31

createConnection

-0.30

 snippetHide

-0.30

 "..\..\

-0.30

íté

-0.30

POSITIVE LOGITS

 Comprometido

0.61

WEBPACK

0.55

хьтан

0.52

 يتيمه

0.49

 CanadaChoose

0.48

 bezeichneter

0.47

 indisponible

0.46

awtextra

0.46

Географиясе

0.46

دانشنامهٔ

0.45

Activations Density 0.050%

requests for information or feedback from users

users, fans, recognize

No Comments

No Known Activations