INDEX

Explanations

phrases indicating time references or sequences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 35-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_35/width_16k/average_l0_120

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.35.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

 MainAxisSize

-0.51

principalColumn

-0.44

shadowOpacity

-0.40

dij

-0.40

 zwar

-0.40

やっと

-0.40

expiresIn

-0.40

unanje

-0.38

 finalmente

-0.38

BagConstraints

-0.38

POSITIVE LOGITS

 outset

1.07

 beginning

0.93

 very

0.88

 inception

0.81

 onset

0.77

beginning

0.74

 moment

0.69

 BEGINNING

0.69

very

0.67

Very

0.66

Activations Density 0.166%

phrases indicating time references or sequences

No Comments

No Known Activations