INDEX

Explanations

content related to scientific reviews and assessments

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.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

lesias

-0.50

Manbalar

-0.44

Datuak

-0.44

 InputDecoration

-0.44

zzleHttp

-0.42

 lenker

-0.41

RectangleBorder

-0.41

ImageContext

-0.41

 aDecoder

-0.41

astify

-0.40

POSITIVE LOGITS

itinéraire

0.48

`}`,

0.45

uxxxx

0.45

енча

0.45

gruntled

0.43

xmlns

0.43

 Heeren

0.42

 renforcé

0.42

 pregă

0.41

 Marino

0.41

Activations Density 0.006%

content related to scientific reviews and assessments

No Comments

No Known Activations