INDEX

Explanations

references to views and viewing concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

期刊论文

-0.67

 bezeichneter

-0.66

Anhalt

-0.62

DialogContent

-0.58

damente

-0.57

UnknownFieldSet

-0.57

autoreleasepool

-0.56

amoja

-0.51

medad

-0.51

ulele

-0.50

POSITIVE LOGITS

ership

0.94

 disambiguazione

0.90

finder

0.80

 AspNetCore

0.77

tiful

0.74

point

0.71

points

0.69

ัฒนา

0.69

 ImageView

0.69

sidemargin

0.68

Activations Density 0.115%

references to views and viewing concepts

No Comments

No Known Activations