INDEX

Explanations

patterns related to text formatting and coding structures

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 overhe

-0.77

BuyableInstoreAndOnline

-0.69

Kin

-0.69

 finances

-0.68

 Exper

-0.67

odynam

-0.66

builders

-0.65

ideos

-0.65

workshop

-0.65

videos

-0.64

POSITIVE LOGITS

 suffix

1.37

 prefix

1.34

 vowel

1.28

 notation

1.28

 denote

1.22

 comma

1.22

 aster

1.21

"+

1.17

 subscript

1.14

 denotes

1.13

Activations Density 1.610%

patterns related to text formatting and coding structures

No Comments

No Known Activations