INDEX

Explanations

terms related to completed surveys

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

Bul

-0.44

 yeux

-0.42

VersionUID

-0.42

Bol

-0.42

oulos

-0.39

her

-0.39

.$.

-0.38

 plea

-0.38

\!

-0.38

 resolution

-0.38

POSITIVE LOGITS

istoitu

0.56

ScopeManager

0.52

ſelves

0.51

 Majefty

0.48

ública

0.46

 Monfieur

0.46

osoba

0.45

 ones

0.44

 mineures

0.44

velocityY

0.43

Activations Density 0.057%

terms related to completed surveys

No Comments

No Known Activations