INDEX

Explanations

relative clauses or phrases that specify or describe nouns

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

EOUS

-0.64

gląda

-0.62

 ratify

-0.60

icionar

-0.60

Ur

-0.59

ciz

-0.58

 aktuelle

-0.58

dė

-0.57

嚆

-0.56

ophan

-0.56

POSITIVE LOGITS

 which

1.14

who

1.07

which

0.99

 ambao

0.90

 joka

0.88

 које

0.84

Which

0.83

 والذي

0.82

 която

0.82

buttonShape

0.81

Activations Density 0.506%

relative clauses or phrases that specify or describe nouns

No Comments

No Known Activations