INDEX

Explanations

the word "is" in various contexts, often indicating a state of being

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 4-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_4/width_16k/average_l0_124

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.4.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

 समीक्षाओं

-0.93

abestanden

-0.83

 المعيارى

-0.76

 Chwiliwch

-0.75

 oprot

-0.72

verwijspagina

-0.67

pyplot

-0.67

Manbalar

-0.64

 pyplot

-0.63

 تكبرها

-0.63

POSITIVE LOGITS

は

1.43

은

1.26

는

1.25

は

1.15

体は

1.08

子は

1.05

さは

1.03

者は

1.03

曲は

1.01

らは

1.01

Activations Density 0.052%

the word "is" in various contexts, often indicating a state of being

No Comments

No Known Activations