INDEX

Explanations

forms of the verb "to be" in various tenses and structures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 39-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_39/width_16k/average_l0_131

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.39.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

MessageTagHelper

-0.39

 xmlns

-0.32

leiter

-0.32

 purpoſe

-0.32

 estime

-0.31

maso

-0.31

DebuggerStep

-0.30

getSeconds

-0.29

LabelTagHelper

-0.28

MeasureSpec

-0.28

POSITIVE LOGITS

 forced

0.82

forced

0.75

 compelled

0.73

 able

0.71

 tempted

0.70

 successful

0.67

InitVars

0.66

 quick

0.65

 rewarded

0.65

 Forced

0.64

Activations Density 0.600%

forms of the verb "to be" in various tenses and structures

No Comments

No Known Activations