INDEX

Explanations

references to programming languages and file loading associated with coding frameworks

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

bin file paths and helpers

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

Autoritní

-0.94

aarrggbb

-0.83

 resourceCulture

-0.79

Personensuche

-0.79

GEBURTSDATUM

-0.78

UnusedPrivate

-0.73

AndEndTag

-0.71

 getItemId

-0.68

 CreateTagHelper

-0.68

ंदीखरीदारी

-0.67

POSITIVE LOGITS

 Dietz

0.52

 Franke

0.51

 Gillespie

0.51

 Metzger

0.48

 Coffey

0.46

<strong>

0.46

 Pearce

0.45

 Pritchard

0.45

 Milne

0.44

Activations Density 0.129%

references to programming languages and file loading associated with coding frameworks

bin file paths and helpers

No Comments

No Known Activations