INDEX

Explanations

references to sections and figures in a document

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

references to sections or figures

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

acu

-0.36

TResult

-0.35

ng

-0.34

 prison

-0.34

UnsafeEnabled

-0.34

NG

-0.34

dru

-0.33

getFrom

-0.33

 driven

-0.32

 Harn

-0.32

POSITIVE LOGITS

RenderAtEndOf

0.71

AsUp

0.63

+#+#

0.60

matchCondition

0.60

httphttps

0.59

DockStyle

0.58

 propOrder

0.56

colhead

0.55

Rptr

0.55

 Obrador

0.54

Activations Density 0.252%

references to sections and figures in a document

references to sections or figures

No Comments

No Known Activations