INDEX

Explanations

terms related to safety regulations and requirements

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

type size, width, between

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

!*\

-0.40

Transkript

-0.39

SuppressLint

-0.37

:+:

-0.35

 wonderland

-0.35

 flipping

-0.35

 passions

-0.34

 unbelievable

-0.34

 anyways

-0.34

 pique

-0.33

POSITIVE LOGITS

 ſind

0.59

 dispositifs

0.55

melidir

0.52

ceptable

0.52

 ModelRenderer

0.50

 Sistemas

0.49

 onOptions

0.48

 nonUne

0.48

 ausreich

0.48

properly

0.48

Activations Density 0.025%

terms related to safety regulations and requirements

type size, width, between

No Comments

No Known Activations