INDEX

Explanations

instances of the word "say" in various forms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 4-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_4/width_16k/average_l0_124

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.4.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

 Staub

-0.76

 Thornton

-0.71

']")

-0.69

 گرد

-0.68

".$_

-0.68

urel

-0.68

lotten

-0.67

وردار

-0.66

Składniki

-0.65

 Thorne

-0.64

POSITIVE LOGITS

say

2.29

SAY

2.23

Say

2.19

Say

2.17

say

2.17

SAY

2.11

 Saying

1.97

 saying

1.91

saying

1.86

Saying

1.79

Activations Density 0.090%

instances of the word "say" in various forms

No Comments

No Known Activations