INDEX

Explanations

variations of the word "occur" in different contexts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Ancho

-0.62

etje

-0.58

 Diagonal

-0.57

Diagonal

-0.57

gine

-0.56

 Anſ

-0.56

tshell

-0.54

Skeleton

-0.54

DT

-0.54

 Alpes

-0.54

POSITIVE LOGITS

 occur

1.27

 occurs

1.19

 Occur

1.12

 occurring

1.08

occur

1.06

 occurred

1.06

 occurrence

1.04

occurs

1.00

 ocurrir

0.98

 ocorrer

0.97

Activations Density 0.016%

variations of the word "occur" in different contexts

No Comments

No Known Activations