INDEX

Explanations

data structures or representations in a specified format

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

 purpoſe

-1.00

 deſt

-0.94

 ſtate

-0.93

 uſe

-0.92

 juſ

-0.92

 pleaſure

-0.91

 beſt

-0.91

 cauſe

-0.90

addCriterion

-0.89

 greateſt

-0.87

POSITIVE LOGITS

0.72

0.66

0.64

0.62

0.59

«

0.53

0.52

0.51

Activations Density 0.296%

data structures or representations in a specified format

No Comments

No Known Activations