INDEX

Explanations

attends to underscore-related tokens from other programming-related tokens

oai_attention-head · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 0-gemmascope-att-16k

Configuration

google/gemma-scope-2b-pt-att/layer_0/width_16k/average_l0_104

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.0.attn.hook_z

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.01

1:0.03

2:0.07

3:0.61

4:0.04

5:0.04

6:0.09

7:0.07

Negative Logits

ſelf

-0.76

 myſelf

-0.76

Efq

-0.75

^(@)

-0.74

IntoConstraints

-0.74

 raiſ

-0.71

 itſelf

-0.68

 ſeveral

-0.68

 purpoſe

-0.67

NUMX

-0.67

POSITIVE LOGITS

0.53

↵

0.45

0.43

0.42

0.41

0.40

0.36

↵↵

0.36

Activations Density 0.057%

attends to underscore-related tokens from other programming-related tokens

No Comments

No Known Activations