INDEX

Explanations

adjectives describing something organized or pleasing

oai_token-act-pair · gpt-3.5-turbo

descriptors related to organization and neatness

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

Downloadha

-0.71

senal

-0.68

ioxide

-0.67

 defenders

-0.67

CVE

-0.66

 Defenders

-0.66

 risked

-0.65

 exerted

-0.64

 authorized

-0.63

POSITIVE LOGITS

nesses

1.09

ness

0.99

liness

0.96

ety

0.90

arity

0.84

 neat

0.83

ilde

0.82

icles

0.80

iness

0.78

 little

0.77

Activations Density 0.026%

adjectives describing something organized or pleasing

descriptors related to organization and neatness

No Comments

No Known Activations