INDEX

Explanations

references to conflicts of interest

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

:✨

-0.34

Grath

-0.34

銳

-0.34

oforte

-0.33

store

-0.33

retail

-0.32

)++;

-0.32

Dtor

-0.31

fol

-0.31

нительно

-0.31

POSITIVE LOGITS

AddTagHelper

0.60

 corrupción

0.59

TargetException

0.58

StoryboardSegue

0.54

 incest

0.54

 Comprometido

0.54

homonymie

0.52

 amizade

0.52

 ujednoznacz

0.52

 שוליים

0.51

Activations Density 0.046%

references to conflicts of interest

No Comments

No Known Activations