INDEX

Explanations

references to "record" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-9b-pt-res/layer_2/width_16k/average_l0_67

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.2.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

ulex

-0.40

iotic

-0.39

uti

-0.39

ugin

-0.39

Hy

-0.38

setTextSize

-0.38

AutoSize

-0.37

 Mühl

-0.36

Hul

-0.36

Wis

-0.35

POSITIVE LOGITS

 record

4.28

record

4.03

 Record

3.97

Record

3.88

 RECORD

3.63

RECORD

3.31

 records

2.94

records

2.77

 Records

2.72

Records

2.66

Activations Density 0.169%

references to "record" in various contexts

No Comments

No Known Activations