INDEX

Explanations

references to individuals or characters in various contexts

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

vzh

-0.41

 geométrico

-0.40

 začet

-0.39

 následu

-0.38

 tijden

-0.37

 colgante

-0.36

 szczegó

-0.36

 integrado

-0.36

 spørs

-0.36

 opleiding

-0.35

POSITIVE LOGITS

 someone

0.82

 createState

0.81

someone

0.80

 somebody

0.80

ſicht

0.78

somebody

0.76

<unused14>

0.76

<unused51>

0.75

<unused43>

0.75

<unused3>

0.75

Activations Density 0.060%

references to individuals or characters in various contexts

No Comments

No Known Activations