INDEX

Explanations

occurrences of the tab character

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

setVerticalGroup

-1.17

 ſind

-0.97



-0.94

 betweenstory

-0.89

Хьажоргаш

-0.88

 queſta

-0.88

RegressionTest

-0.84

 ་་

-0.83

iffance

-0.82

 laſſen

-0.80

POSITIVE LOGITS

The

0.63

the

0.59

0.55

0.52

0.51

0.47

the

0.43

0.42

It

0.41

Activations Density 0.021%

occurrences of the tab character

No Known Activations