INDEX

Explanations

phrases indicating expectations or assumptions about what should happen

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

med

-0.44

stak

-0.35

 señ

-0.33

TS

-0.33

os

-0.32

ote

-0.32

yelidikan

-0.32

-0.31

gard

-0.31

frist

-0.31

POSITIVE LOGITS

elemField

0.69

EDEFAULT

0.65

ViewInit

0.63

本当は

0.56

WriteTagHelper

0.55

MigrationBuilder

0.53

 ब्रेकडाउन

0.53

}{@

0.52

supposed

0.52

 normalement

0.52

Activations Density 0.067%

phrases indicating expectations or assumptions about what should happen

No Comments

No Known Activations