INDEX

Explanations

references to groups of people or entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

ecap

-0.79

 oreilles

-0.67

 wystaw

-0.65

 pleaſure

-0.63

 faſt

-0.63

 nuages

-0.63

SerializedSize

-0.62

these

-0.62

 cdti

-0.62

issau

-0.61

POSITIVE LOGITS

who

1.13

 whom

0.72

ScopeManager

0.69

 же

0.65

 pesky

0.65

 involved

0.62

genen

0.61

enumi

0.60

who

0.60

 else

0.60

Activations Density 0.096%

references to groups of people or entities

No Comments

No Known Activations