INDEX

Explanations

past tense phrases indicating habitual actions or states

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 10-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_10/width_16k/average_l0_113

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.10.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

 referenties

-0.61

 الحره

-0.60

はじめに

-0.53

 ExecuteAsync

-0.52

TIR

-0.49

 aporte

-0.49

 GenerationType

-0.48

 valgt

-0.48

BCA

-0.48

unteer

-0.48

POSITIVE LOGITS

 formerly

0.89

 vroeger

0.81

 früher

0.80

 Formerly

0.73

formerly

0.72

 former

0.72

 autrefois

0.66

 originally

0.65

 einst

0.64

Formerly

0.64

Activations Density 0.194%

past tense phrases indicating habitual actions or states

No Comments

No Known Activations