INDEX

Explanations

elements related to error messages and file references in code

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

code build directives

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

 nadzieję

-0.38

 znamen

-0.35

 Inscrivez

-0.33

 devrez

-0.32

 tepat

-0.32

 chrétiens

-0.31

 trés

-0.31

 institucional

-0.30

printStackTrace

-0.30

 Winaray

-0.30

POSITIVE LOGITS

posedge

0.77

keyframes

0.75

帖最后由

0.74



0.70

 autorytatywna

0.69

عرِّف

0.68

 noDo

0.67

 AssemblyProduct

0.67

IVEREF

0.67

SerializedName

0.66

Activations Density 0.054%

elements related to error messages and file references in code

code build directives

No Comments

No Known Activations