INDEX

Explanations

attends to punctuation tokens marking the end of sentences or clauses from subsequent 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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.01

1:0.03

2:0.02

3:0.06

4:0.04

5:0.03

6:0.04

7:0.73

Negative Logits

 itſelf

-0.93

 myſelf

-0.85

 Monfieur

-0.84

 pleaſure

-0.82

 ―――――

-0.78

 poffe

-0.76

 purpoſe

-0.75

 themſelves

-0.74

 ་་

-0.73

Efq

-0.73

POSITIVE LOGITS

0.45

0.44

en

0.44

0.43

con

0.42

the

0.41

de

0.41

la

0.41

Activations Density 0.018%

attends to punctuation tokens marking the end of sentences or clauses from subsequent tokens

No Comments

No Known Activations