INDEX

Explanations

sections of text that have no activations, indicating it may be looking for formatting or structural cues rather than content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 12-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_12/width_16k/average_l0_82

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.12.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

 محفوظة

-0.50

Soorten

-0.47

associated

-0.46

 فريبيس

-0.45

NameInMap

-0.45

 esperienze

-0.43

Ohr

-0.43

Související

-0.43

相关的

-0.42

égard

-0.42

POSITIVE LOGITS

 '\\;'

0.70

 vuitton

0.66

specialchars

0.65

ônus

0.65

 endblock

0.63

ginx

0.61

yntaxException

0.60

🏻‍♀️

0.60

SPATH

0.58

autique

0.58

Activations Density 0.135%

sections of text that have no activations, indicating it may be looking for formatting or structural cues rather than content

No Comments

No Known Activations