INDEX

Explanations

attends to tokens related to "ak" from tokens designated with "ar" in varied contexts

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

0:0.84

1:0.00

2:0.01

3:0.01

4:0.07

5:0.01

6:0.00

7:0.02

Negative Logits

<eos>

-0.87

-0.59

-0.54

ah

-0.53

el

-0.52

au

-0.51

ad

-0.51

-0.50

ar

-0.50

POSITIVE LOGITS

Efq

1.05

 myſelf

0.93

 Majefty

0.90

 itſelf

0.89

 pleaſure

0.88

 purpoſe

0.86

 houſe

0.84

 ſtate

0.84

 theſe

0.84

^(@)

0.84

Activations Density 23.587%

attends to tokens related to "ak" from tokens designated with "ar" in varied contexts

No Comments

No Known Activations