INDEX

Explanations

phrases related to completeness or totality

oai_token-act-pair · gpt-3.5-turbo

statements that highlight the existence or quality of situations or conditions in a variety of 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

ahime

-0.84

qus

-0.78

edia

-0.78

 angered

-0.69

 Archdemon

-0.66

eln

-0.65

pelling

-0.62

claw

-0.61

vernment

-0.60

ews

-0.60

POSITIVE LOGITS

 except

1.21

 equally

0.93

except

0.88

 imaginable

0.88

Tes

0.80

 alike

0.78

 interchangeable

0.72

winner

0.71

 conceivable

0.67

 equal

0.65

Activations Density 0.411%

phrases related to completeness or totality

statements that highlight the existence or quality of situations or conditions in a variety of contexts

No Comments

No Known Activations