INDEX

Explanations

instances of high numbers or results in a technical context

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

programming code and structure

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

RenderAtEndOf

-0.71

AddTagHelper

-0.69

WebVitals

-0.68

HasForeignKey

-0.65

 CreateTagHelper

-0.63

 ModelExpression

-0.62

tvguidetime

-0.62

PreferredItem

-0.59

IntoConstraints

-0.59

onOptions

-0.59

POSITIVE LOGITS

<eos>

0.38

0.31

0.30

 Neustadt

0.30

金

0.30

法

0.29

zde

0.28

Compile

0.28

pdf

0.28

Activations Density 0.005%

instances of high numbers or results in a technical context

programming code and structure

No Comments

No Known Activations