INDEX

Explanations

numerical sequences or values within the text

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 12-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_12/width_16k/average_l0_130

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.12.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

 pleaſure

-0.69

 Reſ

-0.68

ſelves

-0.67

 purpoſe

-0.61

 Anſ

-0.61

 ſtate

-0.60

 Inſ

-0.59

 ſche

-0.59

ſelf

-0.58

 inſ

-0.58

POSITIVE LOGITS

urilor

0.42

 utafitiHapana

0.42

__(/*!

0.42

Архівовано

0.41

 AssemblyCompany

0.41

 juzgar

0.40

 lecz

0.40

Concerning

0.40

Hochspringen

0.40

 tulang

0.39

Activations Density 0.315%

numerical sequences or values within the text

No Known Activations