INDEX

Explanations

words related to organization and structure

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

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

footed

-0.66

ãĥŃ

-0.62

Rush

-0.60

ãĥ¯

-0.59

mercial

-0.59

 Anglo

-0.58

umbledore

-0.57

asive

-0.57

 Spoon

-0.56

tumblr

-0.56

POSITIVE LOGITS

roth

1.07

opol

0.72

lene

0.70

 unsus

0.69

owa

0.68

ovsky

0.67

lde

0.66

ovic

0.65

mber

0.65

bor

0.65

Activations Density 0.042%

words related to organization and structure

No Comments

No Known Activations