INDEX

Explanations

occurrences of the token "la"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 1-gemmascope-mlp-16k

Configuration

google/gemma-scope-2b-pt-mlp/layer_1/width_16k/average_l0_105

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.1.hook_mlp_out

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

la

-2.75

LA

-1.62

la

-1.41

La

-1.22

ла

-1.16

La

-1.16

LA

-1.08

laa

-1.08

lae

-1.05

las

-0.79

POSITIVE LOGITS

complexContent

0.63

zeg

0.58

 AssemblyTitle

0.57

 MessageBoxIcon

0.57

ipedi

0.55

dropna

0.54

 itſelf

0.53

orations

0.52

 perſon

0.51

 quedarse

0.50

Activations Density 0.067%

occurrences of the token "la"

No Comments

No Known Activations