INDEX

Explanations

phrases in proper English syntax, especially those involving someone mentioning or stating something

oai_token-act-pair · gpt-3.5-turbo

the verb "said" used repeatedly in different contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.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

folios

-0.74

izo

-0.72

ingu

-0.67

 ILCS

-0.66

otin

-0.66

estern

-0.66

LEASE

-0.64

 reunited

-0.63

mol

-0.63

larg

-0.62

POSITIVE LOGITS

WER

0.67

 Donna

0.67

 Angela

0.65

 Hilbert

0.64

 Bulg

0.64

 Charm

0.64

FT

0.64

 according

0.63

 analysts

0.63

 Larry

0.63

Activations Density 0.053%

phrases in proper English syntax, especially those involving someone mentioning or stating something

the verb "said" used repeatedly in different contexts

No Comments

No Known Activations