INDEX

Explanations

references to cutting or shredding actions and related terms

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 state

-0.45

 benessere

-0.42

STATE

-0.42

 Schaefer

-0.40

 câte

-0.40

 encontraban

-0.40

 estado

-0.39

who

-0.39

 synes

-0.39

 răsp

-0.39

POSITIVE LOGITS

 Cutting

1.22

Cutting

1.21

 cutting

1.12

cutting

1.11

cuts

1.08

cut

1.08

CUT

1.05

 cuts

1.02

Cuts

1.02

Cut

1.01

Activations Density 0.333%

references to cutting or shredding actions and related terms

No Comments

No Known Activations