INDEX

Explanations

references to "profile" or similar terms related to user profiles or characteristics

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

 pouvoit

-0.57

<bos>

-0.54

 feroit

-0.53

 verrez

-0.52

 bakteri

-0.52

 mathématiques

-0.51

 démocr

-0.51

 dinheiro

-0.51

 bancaire

-0.50

música

-0.50

POSITIVE LOGITS

 profile

1.22

 Profile

1.13

 PROFILE

1.11

profile

1.09

Profile

1.05

 profiles

1.04

PROFILE

0.96

 Profiles

0.91

Profiles

0.90

getProfile

0.86

Activations Density 0.009%

references to "profile" or similar terms related to user profiles or characteristics

No Comments

No Known Activations