INDEX

Explanations

patterns in whitespace or indentation in code

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

ſelf

-0.84

 itſelf

-0.76

 ſch

-0.74

 Verſ

-0.74

yarnpkg

-0.74

 ſche

-0.69

Efq

-0.69

ſelves

-0.69

 myſelf

-0.68

 AspNetCore

-0.68

POSITIVE LOGITS

UnusedPrivate

0.37

WriteAttribute

0.37

enschappe

0.37

0.31

 GenerationType

0.28

 repente

0.28

0.27

Activations Density 0.007%

patterns in whitespace or indentation in code

No Comments

No Known Activations