INDEX

Explanations

requests for additional information or clarification

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 27-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_27/width_16k/average_l0_118

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.27.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

ніципалі

-0.70

ArrowToggle

-0.66

tonsoft

-0.60

principalTable

-0.59

protoimpl

-0.59

GeneratedCode

-0.58

 مشين

-0.56

Darn

-0.56

imageshack

-0.54

DeleteBehavior

-0.52

POSITIVE LOGITS

 purposes

1.09

 sake

0.85

purposes

0.76

 reasons

0.71

 Purposes

0.70

 purpose

0.60

 Zwecke

0.53

reasons

0.52

来说

0.50

 raisons

0.48

Activations Density 0.809%

requests for additional information or clarification

No Known Activations