INDEX

Explanations

technical or programming-related terminology

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 27-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_27/width_16k/average_l0_118

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.27.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

↵↵

-0.65

-0.64

↵↵↵

-0.43

↵

-0.40

</tbody>

-0.37

}$.

-0.35

↵↵↵↵

-0.34

。

-0.34

<eos>

-0.33

POSITIVE LOGITS

 كومونز

0.91

verwijspagina

0.84

 يتيمه

0.77

ագրություններ

0.75

Chham

0.74

GEBURTSDATUM

0.72

uxxxx

0.72

Datuak

0.72

 lenker

0.71

 kaarangay

0.71

Activations Density 0.776%

technical or programming-related terminology

No Known Activations