INDEX

Explanations

references to specific locations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 22-gemmascope-mlp-16k

Configuration

google/gemma-scope-2b-pt-mlp/layer_22/width_16k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.22.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

Iden

-0.64

 idén

-0.55

as

-0.52

дона

-0.48

ArgsConstructor

-0.46

 telt

-0.46

odw

-0.46

 coque

-0.45

strix

-0.45

 createSlice

-0.45

POSITIVE LOGITS

 AssemblyCulture

0.73

Tikang

0.62

LayoutStyle

0.62

IsContent

0.60

 Diony

0.59

 bVar

0.59

 cherchés

0.57

 Voulez

0.56

hest

0.56

<thead>

0.55

Activations Density 0.938%

references to specific locations

No Comments

No Known Activations