INDEX

Explanations

phrases that refer to reasons or justifications for actions or outcomes

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

unspecified reasons

np_acts-logits-general · gemini-2.5-flash-lite

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 queſta

-0.59

findpost

-0.58

Дереккөздер

-0.56

tagHelperRunner

-0.54

 Paglinawan

-0.53

StoryboardSegue

-0.53

$_(

-0.52

 ddelwed

-0.51

 שוליים

-0.50

 surla

-0.50

POSITIVE LOGITS

 circonstances

0.66

 reasons

0.63

 circunstancias

0.62

 circumstances

0.55

 raisons

0.55

reasons

0.54

 Gründe

0.50

 motivos

0.47

 razones

0.47

Reasons

0.46

Activations Density 0.046%

phrases that refer to reasons or justifications for actions or outcomes

unspecified reasons

No Comments

No Known Activations