INDEX

Explanations

references to personal beliefs and experiences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 36-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_36/width_16k/average_l0_120

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.36.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

 Subsequent

-0.49

 ensuing

-0.47

 subsequent

-0.46

resave

-0.43

 intervening

-0.42

Ongoing

-0.42

 Ongoing

-0.41

 ongoing

-0.40

 Pristupljeno

-0.40

principalTable

-0.39

POSITIVE LOGITS

 formerly

2.61

 previously

2.28

 Formerly

2.23

 autrefois

2.13

formerly

2.11

previously

2.08

Formerly

2.08

 Previously

2.06

 vroeger

2.05

Previously

2.02

Activations Density 0.935%

references to personal beliefs and experiences

No Comments

No Known Activations