INDEX

Explanations

attends to the token "going" from various contexts, while also looking for corresponding "going" tokens that indicate future actions or plans

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-att/layer_0/width_16k/average_l0_61

Prompts (Dashboard)

16,384 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

0:0.03

1:0.04

2:0.02

3:0.11

4:0.34

5:0.07

6:0.04

7:0.06

8:0.04

9:0.02

10:0.02

11:0.04

12:0.04

13:0.03

14:0.03

15:0.01

Negative Logits

 wasn

-0.56

 didn

-0.53

 wouldn

-0.51

 doesn

-0.49

don

-0.48

 shouldn

-0.47

 just

-0.47

 went

-0.46

isn

-0.46

POSITIVE LOGITS

 ainfi

1.00

 Houſe

0.96

ſelf

0.90

 queſta

0.90

 myſelf

0.88

 houſe

0.87

 Anſ

0.86

 Monfieur

0.85

 pleaſure

0.85

 ſche

0.84

Activations Density 0.079%

attends to the token "going" from various contexts, while also looking for corresponding "going" tokens that indicate future actions or plans

No Comments

No Known Activations