INDEX

Explanations

elements related to coding or programming instructions

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-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_22/width_16k/average_l0_72

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.22.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

enumi

-1.01

 Diſ

-1.01

 pleaſure

-0.96

 purpoſe

-0.94

 Esau

-0.91

 CreateTagHelper

-0.91

enumii

-0.90

 Czechoslovakia

-0.90

 الحره

-0.89

 ſche

-0.88

POSITIVE LOGITS

«

0.46

 свой

0.46

its

0.45

runOnUiThread

0.43

 את

0.43

der

0.41

sel

0.41

con

0.41

esa

0.40

 тот

0.40

Activations Density 0.009%

elements related to coding or programming instructions

No Comments

No Known Activations