INDEX

Explanations

phrases indicating varying levels of quality or size

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 5-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_5/width_16k/average_l0_68

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.5.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

 gorą

-0.66

RegressionTest

-0.57

 ALWAYS

-0.57

 Zapo

-0.54

Totally

-0.52

 typeof

-0.52

Always

-0.51

anyak

-0.51

servez

-0.51

 Always

-0.50

POSITIVE LOGITS

 decent

0.88

 decently

0.81

considerable

0.77

 considerable

0.76

 fairly

0.75

 непло

0.74

 sizable

0.74

 Decent

0.70

Abraço

0.70

 sizeable

0.70

Activations Density 0.255%

phrases indicating varying levels of quality or size

No Comments

No Known Activations